Professional Documents
Culture Documents
ASTROPHYSICS AND
SPACE SCIENCE LIBRARY
A SERIES OF BOOKS ON THE RECENT DEVELOPMENTS
Editorial Board
B. M. MCCORMAC, Lockheed Palo Alto Research Laboratory, Palo Alto, Calif., U.S.A.
VOLUME 54
PROCEEDINGS
IMAGE
PROCESSING TECHNIQUES
IN ASTRONOMY
PROCEEDINGS OF A CONFERENCE HELD IN UTRECHT
ON MARCH 25-27, 1975
Edited by
INTRODUCTION XI
PART 1
WHAT INFORMATION DO WE NEED, FOR WHICH ASTRONOMICAL PROBLEM?
K. Aa. Strand
ASTROMETRY 3
J. Borgman
PHOTOMETRY 11
W. C. Livingston
THE PHOTOMETRY OF EXTENDED OBJECTS: DIODE ARRAY DETECTORS AND
INTERACTIVE DATA PROCESSING 21
K. o. Wright
TRENDS IN ASTRONOMICAL SPECTROSCOPY 31
PART 2
ACQUISITION AND STORAGE 49
T. de Groot
INFORMATION RETRIEVAL BY MICRODENSITOMETRY 51
P. Connes
FOURIER TECHNIQUES 57
A. Boksenberg
TELEVISION DETECTOR TECHNIQUES 59
J. G. Timothy
PHOTON-COUNTING DETECTOR ARRAYS BASED ON MICROCHANNEL ARRAY
PLATES 85
VI TABLE OF CONTENTS
D. Dravins
A SELF-SCANNED SILICON DIODE ARRAY FOR ASTRONOMICAL
PHOTOMETRY 97
W. M. Burton
IMAGE DETECTORS FOR RECORDING ASTRONOMICAL ULTRAVIOLET
SPECTRA 103
C. D. Mackay
INTEGRATING TV SYSTEM DEVELOPMENT AT CAMBRIDGE 125
T. M. Karnperman
SEC VIDICON SYSTEM FOR A BALLOON ULTRAVIOLET STELLAR
SPECTROMETER 127
PART 3
PROCESSING HARDWARE 157
J. S. Newcomb
FOUR YEARS EXPERIENCE IN APMS STAR PLATE PROCESSING -
RESULTS AND FUTURE PLANS 159
E. W. Dennison
THE GRANT MEASURING ENGINE 199
TABLE OF CONTENTS VII
K. Aa. Strand
THE U.S. NAVAL OBSERVATORY'S AUTOMATIC MEASURING MACHINES 209
J. L. Hershey
THE TWO-COORDINATE GRANT MACHINE IN LONG-FOCUS ASTROMETRY 211
R. M. West
THE ESO S-3000 MEASURING MACHINE 223
F. Launay
PHOTOELECTRIC SPECTROSCOPIC PLATE MEASURING MACHINE BUILT
AT THE MEUDON OBSERVATORY 265
PART 4
SOFTWARE TECHNIQUES 269
D. A. Klinglesrnith
SOFTWARE TECHNIQUES FOR IMAGE PROCESSING 271
H. Nieuwenhuijzen
PROBLEMS OF MACHINE INDEPENDENCY AND INTERACTIVE MATTERS 285
W. N. Brouw
APERTURE SYNTHESIS 301
VIII TABLE OF CONTENTS
E. H~g
EVALUATION OF MASTER SCANS OF PHOTOGRAPHIC STELLAR IMAGES 323
PART 5
APPLICATIONS TO ASTRONOMICAL PROBLEMS 341
J. Domanget
SOME PRELIMINARY RESULTS OBTAINED WITH A TV CAMERA IN THE
FIELD OF THE OBSERVATION OF VISUAL ~UBLE STARS 343
C. de Jager
H. Nieuwenhuijzen
editors
PAR T
K. Aa. Strand
U. S. Naval Observatory
Washington, D. C,
INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 3·10. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht·Holland.
4 K. Aa. STRAND
REFERENCE SYSTEMS
ASTROGRAPHIC CATALOGS
Since most of the stars for -the program were initially selec-
ted from the Lowell proper motion survey because of the availability
of identification charts, parallaxes have been obtained thus far
for only a few stars fainter than ~ = 16. Fainter stars selected
8 K. Aa. STRAND
from the Luyten 48-inch Palomar Schmidt proper motion survey, are
now on the program. With exposure times of 50 minutes, a photo-
graphic or visual magnitude of 18 is reached.
At the present time, the only other reflectors being used for
parallax work are the one meter aperture Turin astrometric reflec-
tor recently installed and the 98-inch Isaac Newton Telescope.
With the latter telescope, because of the limited availability of
time for such work, a small program restricted to some 20 faint
stars between photographic magnitUde 16 and 18.5 is being carried
out (Murray 1974).
The bulk of the observations are still carried out with the
conventional micrometer, but there are only 3 institutions in the
United states where such observations are made regularly, a few
more in Europe, and only one in the southern hemisphere. Clearly
this combined effort is insufficient to keep track of some 70,000
known binaries of which approximately 1 percent have determined
orbits.
RADIO ASTROMETRY
CONCLUSION
REFERENCES
J. Borgman
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 11:20. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
12 J. BORGMAN
50
-<
~
-<
~
c 20
0
+-'
::J
'010
tf)
Q)
L-
a 5
L-
+-'
U
Q)
0..
If)
2
0.1 10 100 1000
A(jl)
classical photometry (3000 11. < A < 9000 11., see the
comprehensive review by Golay, 1974) but will include
the ultraviolet and infrared regions. The spectral
resolution limits are chosen to include on the one side
wide-band photometry in a full atmospheric window (e.g.
8 - 13~) but to stop at the higher resolution where
narrow-band photometry yields equivalent widths of
individual hydrogen lines. In between lies the realm of
photometry with observable quantities as schematically
illustrated in Fig. 1.
c 2
Ol (8.1-12.2 )
a
E
x
Ql (40 -100)
"0
C 0
'-
::J (9 -V)
~ -1
0
u
-2 (.15 -.25)
-3
10 2 103
Temperature (K)
system, etc.
-linear combinations of colour indices to measure the
blanketing by groups of lines due to a single chemical
element; e.g. the depression caused by the crowding of
hydrogen lines in front of the Balmer discontinuity:
o in the Groningen system, [B - ~ in Walraven's (1960)
system.
1550A
·f
6 em
u
w
0
"0
If)
°co
lD
I
"If)
If)
°co
lD
I
"0
0
Om
I.D
I
" If)
0
Om
lD
I
5h 38 m 37 m 36m RA
CONCLUSION
REFERENCES
W. C. Livingston
The astronomer now possesses two new tools which assist him in
the deduction of the physical conditions in extended objects.
The first is a piece of hardware, the sol id-state diode-array
detector, which wi II I ikely supplant most vacuum devices within
a few years. Diode-arrays offer fundamental photometric advan-
tages over television type transducers (for example, SEC vidicons)
and even other photoemissive tubes (including electronographic
types) because they have a fixed geometry and a determined (and
fixed) transfer characteristic. Surface photometry equal in pre-
cision to that afforded by the photomultipl ier tube becomes fea-
sible. The resulting 2-D arrays of data present some difficulties
in interpretation, but their digestion is aided by the second new
tool, primari Iy a technique, which we shall call "interactive data
processing". Computer generated displays of data fields are pre-
sented in pictorial or graphical form. The display is not in hard
copy but by a CRT display which is under on-I ine computer control.
The astronomer then must interact with his data within-seconds,
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 21-30. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
22 W. C. LIVINGSTON
I. 0 lODE-ARRAYS
~
CDC 6400 Computer Disc Drives
6SS36 60-bit. words 3.S x 10 7 60-bit words
(3.9 x 10 6 bits) (2.1 x 10 9 bits, maximum
1
transfer rate about
S x 10 6 bps)
Computer-computer
Interface, maximum
:\
7-track tape @75ips transfer rate 1 x 10 6
(3.6 x 105 bps) l2-bi t words per second
(1.2 x 10 7 bps)
9-track tape @37.5ips
(2.4 x 105 bps) \ Varian 1
V74 Computer
Disc Drive
2.5 x 10 6 l6-bit words
(4.0 x 10 7 bits, maximum
Icard reader @250cpm t--l 32768 16-bit words ~ transfer rate
(S.2 x lOS bits) 1.6 x 10 6 bps)
Tektronix 4010 semiconductor memory
Interactive Graphics
Terminal
1023 x 780 raster
250 vectors per second
V microprogrammed
REFERENCES
H~g, E. and Wiskott, D.: 1974, European Southern Obs. Tech. Re-
port No.5.
K. O. Wright
INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 31-48. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
32 K. O. WRIGHT
STELLAR EVOLUTION
NUCLEOGENESIS
STELLAR ATMOSPHERES
STELLAR CHROMOSPHERES
Among the developments of the past few years has been the
acceptance of the fact that many stars have extensive chromospheres.
By analogy with the sun and observations made at solar eclipses,
this phenomenon was to be expected but the observational e~idence
was difficult to obtain. The s Aurigae eclipsing systems, with
a giant primary star that eclipses a small hot companion, have been
shown to have chromospheres since lines arising from low-excitation
levels are greatly enhanced close to eclipse and decrease in
strength rapidly as the line of sight of the hot star passes beyond
the outer atmosphere of the giant primary star. Some of the best-
analyzed data are those for s Aurigae made by Wilson and Abt (1954),
though similar observations for 31 and 32 Cygni have been obtained
at Victoria for a number of years .. We have finally completed a
TRENDS IN ASTRONOMICAL SPECTROSCOPY 41
computer program that separates the spectra of the two stars and,
after further analysis, the final tracing shown in Figure 1 is,
we believe, the pure chromo spheric spectrum of the K star.
such as T Tauri and the RV Tauri stars. The source of the envel-
ope is of considerable interest. Is it contracting or expanding?
When polarization effects are observed, the explanation is usually
given in terms of stellar envelopes since it is thought that, in
this case, they arise from electron scattering of dust particles.
More observations are needed.
DOUBLE STARS
-lZ
-10; -35 -4; -34 +2; -37 -1; -41
i:
-10
3 4 5 8
-25 ~3
~
+80 -12
-4 +7 +74 +2
-61
+66 -106
oro!
-32
+3 .~)
+49
-11
2 +77 7
-60
0'89 -112
-17; -39
0'04
;fa
-13;-36
+74
+13
.~+76
-96~+65 Ui
-69 -22 -40
-25
0'82 10 -61 9
-80
-22; -,9 -18
+97
Ha VELOCITIES
v • I• .. 0'43
-31 ;-8
0·25
-37 ;-3
0'12
-30; -9
emission absorption
PHASE
M- VELOCITY -0
VARIABLE STARS
able conditions.
One star that is not typical, but that has attracted much
attention is FG Sagittae (Kraft, 1974) where the spectrum has
changed markedly in the past fifteen years and the luminosity has
increased by four magnitudes during the last fifty years. It may
be the central star in a planetary nebula; it is a cool star and
seems to be changing from a normal F-type giant to a barium- or
even an S-type star, cooling by 250 0 per year and increasing the
abundance of the heavy elements in its atmosphere. Studies of
stars such as FG Sagittae must begin as soon as the peculiarities
are discovered since from such studies, our information about the
processes of evolution and nucleosynthesis will be greatly in-
creased.
EXTRA-GALACTIC ASTRONOMY
Many large-scale features in our own Galaxy have not yet been
related to structures in other galaxies, but in our own neighbour-
hood we sometimes cannot see the forest for the trees. In extra-
galactic systems, their faintness and small size, which are direct-
ly related to their distances, make observations difficult. How-
ever Sandage is finally obtaining a consistent value for the Hubble
constant, which is fundamental in deriving distances of galaxies
from their red shifts. Much more work remains to be done on the
quasi-stellar objects, though these programs have been given high
priority at the large telescopes in order to complete surveys to
18m. The different velocity systems found between emission and
absorption lines should not be considered unreasonable since the
lines may not be formed at the same place, and especially since
similar differences have been observed even in stellar atmospheres;
however the explanation for these differences must be found. The
46 K. O. WRIGHT
SUMMARY
REFERENCES
T. de Groot
I I ®P (convolution)
n o
J J T
n o
de Jager!Nieuwenhuijzen (eds.), Irruzge Processing Techniques in Astronomy. 51-55. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
52 T. DEGROOT
Microdensitometer non-linearity
The only quantity that is allowed to influence the transmittance
during a scan is the grain density within the boundaries of the
scanning spot. The least trivial causes of non-linearity, i.e. of
a variable system transfer function, are defocussing during the
scan and coherence effects. Others, such as "flare" and stray
light are controllable and can be kept within acceptable limits.
The danger of defocussing is always present with films and
large size plates. It is only recently that autofocus systems
have been incorporated, to continually control the position of the
exact focus with respect to the emulsion layer (cf. the contribution
of J.R.W. Heintze).
54
>
REFERENCES
I. J. Becherer and G.B. Parrent Jr., J. Opt. Soc. Am. ~Z, 1479,
1967.
2. M.J. Beran and G.B. Parrent Jr., Theory of Partial Coherence,
1964.
3. P. Bouchareine, Optica Acta 11, 585, 1970.
4. J.W. Brault and O.R. White, Astron. and Astrophys., l~, 169,
1971.
INFORMA TION RETRIEVAL BY MICRODENSITOMETRY 55
5. D. Grimes, J. Opt. Soc. Am., 21. 1263, 1971.
6. E.H. Linfoot, Fourier Hethods-In Optical Image Evaluation,
1964.
7. G.O. Reynolds and A.E. Smith, Appl. Optics, ~~, 1259, 1973.
8. R.E. Swing, J. Opt. Soc. Am., 60, 1551A, 1970.
9. R.E. Swing, J. Opt. Soc. Am., ~~, 199, 1972.
10. I. Weingartner, Optik, 32, 5, 308, 1971.
11. I. Weingartner, W. Hira~de, E. Menzel, Optik, ~~, 1, 53, 1971.
12. J.e. Dainty. Optica Acta, 18. 795, 1971-.----
== .
FOURIER TECHNIQUES
P. Connes
A. Boksenberg
10 GENERAL METHODS
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 59-78. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
60 A. BOKSENBERG
PHOTON
I
~ ~ ~
rPHOTOCATHOOE ANO
ELECTRON OPTICS
I IMAGE CONVERTER/I
INTENSIFIER
I PHOTOGRAPHIC
EMULSION
I
I
I PHOTOCATHODE AND I
ELECTRON OPTICS
1
rELECTRO-
1
GRAPHIC
1 I
.-
ELECTRON I •
MICRO-CHANNEL
PLATE ELECTRON
I
•
EMULSION PHOSPHOR MULTIPLIER
, __ :1 __ ,
I FURTHER
: STAGES IN
I
I
I ELECTR(»J
PHOSPHOR
I
L.f~S.ffPE _J
l
r
SOLID STATE
DETECTOR ARRAY
I
•
OPTICAL
COUPLNG I
1
I +
I
PHOTOGRAPHIC
ELECTRICALLY
INDIVIDUALLY
WIRED SILICON MULTIPLEXED r EMULSION
+
DIODES SILICON DIODES,
CHARGE - COUPLED
DEVICES I TELEVISION CAMERA
I
STORAGE TARGET
DEVICES GT:5 1
I
PHOTOCATHODE AND
VIDICON, ELECTRON OPTI CS
LEAD OXIDE VIDICON,
I
SILICON VIDICON,
SOLID STATE
•
DETECTOR ARRAY
l
I
IMAGE
STORAGE TARGET DEVICES DISSECTOR
GT - 10 GT - 100 GT -1000
SECONDARY - ELECTRON -
IMAGE ORTHICON,
ELECTRON BOMBARDED
I MAGE ISOCON CONDUCTION SILICON TUBE
TUBE (SEC) (EBS, SIT)
I SUN SENSORS
\ \\ /I / \ \\ /I /
ECHELLE BLAZE , 6 S 10 ·8·6·, ECHELLE ·BLAZE ,68 1 0 ·8 ·6,
---k::
PHOTOCATHODE
UV PHOTOCATHODE
SEC TARGET
::: ~
PROXIMITY
FOCUSSED CATHODOLUMINESCENT
CONVERTER PHQSPHER
/
z
MgF WINDOW
~
~
o
~
'"
Experimental I points)
1·0 UV Proxlcon WX -31999
0·9 Cs-Tel MgF2 - - - - - - - t - - - t - - (
0·8
Spotial frequency: 5·5 Ip/mm.
0,7 Wp
0·22 eV
0·6
0'5
M
0·4
0·3
0'2
1'01L-__~~~~J-____~____-L__-L__L-~
1 2 3 4
Photocathode Potential IkV)
M
L.-- ~
/'
0'5
I
1\
---
7'0lp/mm
,,""
0'4
j-e.
/V~
~~ ::7/V
0·3
O· 2 ......,
~
O· 1
0 3200 36 00
1200 1600 2000 2400 2800
Wavelength Ii)
Other advantages of the system arise from the use of the on-
line digital memory and computer: (i) storage capacity, in
TELEVISION DETECTOR TECHNIQUES 71
The lens for coupling the intensifier and camera tube has
been specially designed by Professor Wynne of Imperial College,
London. It is optimised in the blue, has an aperture ratio f/2
and, at a working magnification of 0.59, gives respective fields
at the intensifier and camera tube of 35 rnm and 21 rnm. Taking
into account the quantum efficiency of the first photocathode of
the intensifier, a blue light gain of 10 7 means that on average
about 10 8 photons are generated at the final phosphor screen for
each primary photoelectron. About 1% of these are transferred by
the coupling lens to the Plumbicon tube where about 70% of those
arriving are effective in generating a signal electron in the
target. Thus, on average, an individual Srimary photoelectron
results in the generation of about 7 x 10 electrons in the
Plumbicon target. This gives an abundant video signal for the
purposes of the processing functions outlined above. Compression
of weak signals is overcome by means of bias illumination, which
improves beam acceptance by building up the target potential to
about 1 volt between readouts.
~
o'"z
§
("l
EMI 4 STAGE CASCADE INTENSIFIER
SHIELDS
Cl:>;l
;j
@
~
to
ei:lj
'"
Fig. 8. Sectional view of the image intensifier, coupling lens and television camera head
for the VCL Image Photon Counting System.
-l
W
74 A. BOKSENBERG
Figo 9. Raw video output from <1% of total frame showing single
photon events in an isometric display (IPCS).
1O~--~~----------------------------------------------'
9 ----I
AMPLIFIER AND
UJ RESIDUAL NOISE
~ INTEGRAL DISTRIBUTION
ct:
f--
!Z7 PHOTON EVENTS
6U6
Z
oct:
~
~4
15
iE 3
UJ
~2
!«
....I
l!:!1
°O~--~----~----~----'~--~----~----~--~8----~----~10
CENTERING lOOfC
IN
3000
1000
6000
OUT
3000
1000
~
.-
~
:::>
o
u~
Fig. 12. Spectrum of a faint, normal galaxy obtained with the IPes on the Palomar 200-inch ?>
t:Il
telescope at 250~ rom- 1 • The top trace is object + sky, the middle is sky alone and o
:>"i
the bottom is the sky-subtracted spectrum of the galaxy. m
'"
Z
t:Il
m
::0
CJ
NGC 4151
§
\!:l
~
t:l
t!'l
~..,
t' o
~
....
r, t;l
g
~
o
C
t!'l
'"
!..•
:.l
.r
":
..... ...... .... ...... .... ....., .... .......... - ...... -..... ...", ...... .... - ,"('L'~ ~OMS~i.:·......
m.. ...... - ...... .....
Fig. 13. Spectrum of NGC 4151 obtained with the IPCS on the INT at 30R mm- 1 • The upper trace
is the sky-subtracted spectrum of the object. The lower is a vertical expansion of the
upper to bring up the fine detail.
-..J
-..J
78 A. BOKSENBERG
REFERENCES
Beaver, E.A. and McIlwain, C.E: 1971, Rev. Sci. Instr., 42,
1321.
Beyer, R.R. and Goetze, G.W: 1966, Adv. in Electronics and
Electron Physics, 22A, 241.
Boksenberg, A: 1972, Proc. of ESO/CERN Conference:'Auxi1iary
Instrumentation for Large Optical Telescopes', ESO, Geneva,
295.
Grant, J.M: 1966, Proc. I.E.E.E •• 54, 801.
Needham, M.J. and Thumwood. R.F: 1969. Adv. in Electronics and
Electron,Physics, 28A, 129.
Robinson, L.B. and Hamp1er, E.J: 1972, Pub1. Astr. Soc. Pac.
84, 161.
Spicer, W.E: private communication.
Taft, E. and Apker, L: 1953, J. O.S. A. 43, 81.
ON LINE DIGITAL CORRELATION OF PHOTON COUNTING TV
IMAGES FOR STELLAR INTERFEROMETRY
I. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 79-84. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
80 A. BLAZIT ET AL.
1-----------.1)( Co u n t'er x
~ ideo
10 ca\c.ulat"or
(c.ol'l\pvter)
3. PROGRAMMED PROCESSING
t=o(N(N-I)L /L.
w 1
S/N=R~
For low levels of light,R is proportionnal to the num-
ber of photons per frame (2,3):
S/N=eVt/t>I..
In that case SIN does not depend on the number of photons
per frame.For bright images,the photons overlap and ~ld
grey levels.We then adjust the threshold in the interface
to see just the peaks of the speckles.The AC is then dis-
torted,but it is weak in our case.
I;
.slow F•• r- M'c..ro r"D9r.""".~
C.O"Pllt.~. c.o"'pul"t-r$ (.""' pd I-c.rs.
lams -----
~----~----~--------------~~----
ti Q rl wi red
Cor .. e I .. hi,.·
o
50 -100 -'ISO 200 Nph
4. HARD-WIRED AUTOCORRELATOR
Pho\'"on!>
'00 rdi n Q res
I
Int"egrat"ion
me-mory
5. ACTUAL RESULTS
-2
AUR resolved as double star,separation =5.7 10 "
This star has been first resolved optically by Ander-
son with his interferometer in I922:it has been previously
known as a spectral binary with a period of 104 days.
The ACs done corroborate the results obtained by the
FT method.
ORI possibly resolved:one can see a very weakly contra-
sted peak_~us't below the central peak. The separation would
be 3.9 10 ".This is confirmed by the optical FTs
84 A. BLAZIT ET AL.
.
made from the same video sequence.This feature and several
others need further observations before a conclusion
can be made.
REFERENCES
I. Labeyrie,Bonneau,Stachnic,Gesari,ApJ,I94,I47,I974
2. Dainty,Mon.Not.R.astr.Soc.,I69,63I,I9~
3. Roddier,preprint. ---
4. J.Max,Methodes et techniques de traitement du signal.
5. D. Bonneau,These.
PHOTON-COUNTING DETECTOR ARRAYS BASED ON MICROCHANNEL ARRAY
PLATES
J.G. Timothy
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), lnuzge Processing Techniques in Astronomy. 85-90. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht·Holland,
86 J. G. TIMOTHY
PROXIMITY - FOCUSED
DISCRETE - ANODE ARRAY
PROXIMITY - FOCUSED
DISCRETE - ANODE COINCIDENCE
ARRAY
t
We hav evaluated the performance of cascaded MCP's at EUV
wavelengths and have investigated techniques to eliminate the
problems of ion-feedback. The angled-field MCP produced by Gal-
ileo, Inc., shows promise but suffers from problems of gain re-
duction because of the accumulation of space-charge on the in-
sulating strips in the channel wall used to establish the angled
electrostatic field. 5 We are currently evaluating the perform-
ance of MCP's supplied by Mullard, Ltd., which employ curved
microchannels to inhibit ion-feedback. The initial results of
these tests have proved highly successful.
REFERENCES
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 91-96. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
92 T. B. McCORD ET AL.
dard vidicons,
using magnetic
focus and deflec-
tion, and is cooled
with dry ice.
The digital
scan configuration
is 256 horizontal
elements by 250
lines. A frame
of 64,000 pixels
Figure 1 is read in about
4.5 seconds in 70wm steps. Output from the system is
displayed on a CRT monitor, which is refreshed by a
scan converter memory (Brooke, 1975) and also is recorded
on standard 9-track magnetic tape.
The vidio signal is encoded to 12 bits precision
yielding data numbers (DN) ranging from 0 to 4095.
Usually system gain is set so that target saturation
(~106 charge carrier/pixel) is about 10% above DN=4095.
Target preparation is required in order to obtain
reproducible photometric images, and the preparation
must be carried out in exactly the same way for each
image. An image recording routine goes as follows:
(1) target is saturated by exposing i t to bright light,
(2) target is erased 10 times, (3) cathode heater is
turned off to prevent target backlighting, (4) shutter is
opened to begin exposure, (5) shutter is closed, (6)
cathode heater is turned on (a pause is allowed while
he2.ter warms up), (7) target is read out once.
The target-to-cathode-potential is adjusted between
erase and read to place the image signal on a pedestal
signal equal in amplitude to about 10% target saturation.
Backlighting of the infrared responsive silicon
target by the cathode heater contributes a non-spatially-
uniform signal on the order of 1% of saturation in a
minute of warm operation at 6.5 volts. The exact amount
depends on the tube used.
The amount of signal remaining on the target with
each erase cycle decreases rapidly and is reproducible
from erase sequence to sequence to at least within 0.1%
of saturation signal level.
Two types of calibration images are required to
produce an image which is photometrically precise. A
dark field (OF) is made by carrying out the procedures
for making an exposure except for opening the shutter.
A flat field (FF) is made by imaging a uniformly illu-
minated field. The flat field must be made using any
filters in the optical train and the same f/ratio beam
THE MIT SILICON VIDICON IMAGING SYSTEM AT THE TELESCOPE 93
as used for the raw data image. The DF frame is sub-
tracted from both the object and the FF frame. The ob-
ject is divided by the FF frame.
A contour plot and an intensity traverse for a
typical DF and a FF when the full vidicon target is
scanned is shown in figure 2. The frequency distribu-
"i
'"w ,
'"z" ',
CD "
..
... ,~
c(
o ~
DARK FIELD
Figure 2 FLAT FIELD
232.289 ,883
.863 .02)%
.022%
ure 4) show pixel r.m.s.
232.064 .847 .021% deviation of about 0.04%
231.967 .81Z .020% of full scale (4095DN) or
232.067 .859 . 021%
about 1000 charge carriers .
232.116 .849 .021%
The expected r.m.s. devia-
Figure 3 tion for this sample ~106
94 T. B. McCORD ET AL.
events (charge
carriers creation
in the target)
plus 400 charge
carriers per am-
plifier noise is
about 1500 charge
carriers includ-
DATA NUMBER ing the effects
of the division.
"000 [ The difference
'0000, could well be
I
within the uncer-
tainty of our cal-
cula tion. Thus
the imaging sys-
tem is apparently
photon noise lim-
ited above the
effects of a ~500
PIXEL IQ PIXEL
FRAME NO. MEAN DATA NO. ~ DEVIATION ~ FULL SCALE charge carrier
preamplifier noise.
2366.628 I.B .044% The response
2367.127 1.7 .042%
of the system is
2365.800 1.6 .039%
linear (figure 5),
i.e. y=l.OO. The
4 2363.617 '.8 .044%
Figure 5
REFERENCES
Dainis Dravins
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 97-102. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
98 DAINIS DRA VINS
Fig.1. (a)The diode camera (A) is here equipped with a reflex find-
er to easily locate features. At B is a rack unit with external
controls. The signal can be inspected on the adjacent oscilloscope.
(b)Front end of the camera; the rectangular chip with the diode
array is seen in center. Diameter of inner circle = 25 mm.
r- ---------T----------" r-----------------,
I I I
CAMERA RET ICON lC 600 I CONTROL UNIT lMS 102 I CABLE BOM I COMPUHR HP 2116 B
:
I
----~I I
AMPl.
I I
I I
IL I
____________ __________ J L _________________ ~
2. PHOTOMETRIC PROPERTIES
Dark Signal
signal FIXED
t PATTERN
Saturation-level
ruu ODD-EVEN
ASYMMETRY
~UU ...
"''"'UUU~ r""''' --'''''U~
~.' ~ ~,.,
log
signal
3
Saturation-level -+
log illumination
3. ASTRONOMICAL APPLICATIONS
engine has been built at Lund Observatory which can use this diode
array as a detector. The array system can be moved from the solar
telescope to the laboratory photometer in a matter of minutes.
"V\~ /~YVy
3. Atm H20>" 5888.703
4. , 5888.898
5. Atm H20>' 5889.065
6. , 5889.370
7. Atm H20), 5889 .637
8. Iia I DZ >. 5889 .973
9. Co I. Fe I A 5890.495
(
1O.Atm >. 5890.734
11.Atm? >.. 5890.909
12.Atm H2 0. Fe I ). 5891.118
13.Fe II >. 5691.361
14.Atm>' 5891.500
15.Atm H20 ). 58~1.660
16.Fe I , 5891.887
17.Atm? >. 5692.055
lB.Atm? }. 5692.277
19.Atm H20 }. 5892.397
20.Fe I ). 5692.700
21.Ni I }. 5692.883
22.Atm H20 A 5693.045
ACKNOWLEDGEMENTS
BIBLIOGRAPHY
W.M. Burton
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 103-108. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
104 W. M. BURTON
Output
r- Fiber
OpllC
Phosphor
a b
1200 ~ 3000 ~
3. ELECTRONOGRAPHIC DETECTORS
Plan
ov
-ic=========j Solenoid
-20kV ov
' - -_ _ _ _ _ _ _ _ _- ' Solenoid
ACKNOWLEDGEMENTS
REFERENCES
Burton, W.M. and Powell, B.A: 1973 Appl. Opt. 12, 87.
Burton, W.M., Reay, N.K., Shenton, D.B. and Wilson, R: New
Techniques in Space Astronomy (IAU Symposium No. 41), Labuhn
and Lust Eds. (Reidel, Holland 1971) p. 304-312.
Carruthers, G.C: 1972, Adv. Electronics and Electron Physics,
33B, 895.
MESURE DES CONTOURS DE RAIES PAR LES TECHNIQUES
DE TELEVISION
de Jager/Nieuwenhuijzen (eds.), IT1II1ge Processing Techniques in Astronomy. 109-116. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing ,company, Dordrecht-Holland.
110 CH. FEHRENBACH AND A. VIN
-~"~'~~
~.,;....",_....;l..;;.tli)j,,!~JiJ~
',,+,!,:Spl~t .
Figure 2. - Scan TV
Himlirt.
2"'S12~1Zbits.
,--
j,r,NAL I---
Tiltre Integrdtevr lchanlillonnevr r- AbC
.... f::.---
VI/:,f.O ~
L--- 1
Addilionneur
JIGHAL ~
filtre t- IntigrdlelK t- £chantillonnelK t- AIJC
VlM.D [2
~
'--
Figure 5. - 32 UMa
.... -j
.1
i :!;' j: '
TV./hHI
i -b-iL[LL:EI_Eti~l_ ' AMP .....
INTRODUCTION
EMULSION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 117-124. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
118 J. RING AND S. P. WORSWICK
THE MICRODENSITOMETER
LINEARITY
MICRODENSITOMETER NOISE
·10
·05
./
/'
0
....-:::::.
/'
- -'- ..... - -- -EMULSION
·IS NOISE
<Yo 20 MICRON SLIT
-~MICRO-
./ DENSITOMETER
/'~
NOISE
·10 /'
/'
/' -·-·-2Y. RESIDUAL
/' PHOTOCATHOD~
NON-UNIFORMITIES
/'
.oS
- - - IX RESIDUAL
-- PHOTOCATHODE
NON-UNIFORMiTIES
0 - - - _112% RESIDUAL
·15 PHOTOCATHODE
<YO 40 Mle RON sLir NON-UNIFORMITIES
/'
.,./
·10
./"
.,./
.,./
,.-/
·05
.,./
./
-------
./"
-~-7
0
2 3 -----------
4 5
DENSITY
6
COMPUTER PROCESSING
REFERENCES
C.D. Mackay
* This work has been supported 1ll part by the Royal Society and
the Radcliffe Trustees.
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 125-126. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
126 C.D.MACKAY
T.M. Kamperman
1. INTRODUCTION
3. DETECTOR SYSTEM
4. TESTING HARDWARE
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 127-134. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
128 T. M. KAMPERMAN
SPECTROMETER
PIVOT CENTER a ENTRANCE STOP
~~~~~===t[C========C=~-~==~-==~M1~tt_f_.
P ---~"CHRO" ."ROR~
POSITION
SENSOR
Fig.2. BUSS telescope ( NASA- JSC ).
SEC VIDICON SYSTEM FOR A BALLOON ULTRAVIOLET STELLAR SPECTROMETER 129
The Image dissector tube used by NASA has the same disadvantages
for spectrometry as our 5-59 UV Spectrophotometer (de Jager et
al., 1974) in the European TDI-A satellite: relatively long
integration times are required as the system lacks an integrating
element. In the collaboration program we chose for an integrating
television-type detection system behind an echelle-type spectro-
graph.
The goals for the system currently in development at our
laboratory are the following:
- Wavelength region: 200 - 340 nm. The short wavelength limit is
imposed by the remaining atmosp~ere at float altitude of 40 km.
A spectral resolution of 3 x 10 to enable studies of the
dynamic properties of stellar atmospheres, abundancies of rare
elements, their isotopes, and the interstellar medium.
Observation of m = 5 stars (of type B ) with 3% noise in IS
. v 0
m~nutes.
TELESCOPE
f = 3000mm
F / 7.5 I'
I'
1\
I \ CATADIOPTRIC
I IMAGING
I
I SYSTEM
I f =510mm F/4.5
I
I
I
~///~ ECHELLE 79l/mm \
DICHROIC
MIRROR
OFF-AXIS
PARABOLOIDAL
MIRROR
f =533mm F/7.5
:-l
~
FUSED SILICA PRISM ~
(angle 15°)
( reflecting backsurface) ~
Fig. 3. Optical lay-out spectrograph. :;tI
s:::
;I>
z
SEC VIDICON SYSTEM FOR A BALLOON ULTRA VIOLET STELLAR SPECTROMETER 131
3. DETECTOR SYSTEM
4. TESTING HARDWARE
REFERENCES
Kondo, Y., Giuli, R.T., Modisette, J.L. and Rydgren, A.E.: 1972,
Astrophy. Journ. 176, 153.
Jager, C. de, Hoekstra, R., Hucht, K.A. van der, Kamperman, T.M.,
Lamers, H.J., Hammerschlag, A., Werner, W., and Emming, J.G.:
1974, Astroph. and Space Science~, 207.
Flory, L.E. ,Pike, W.S., Morgan, J.M. and Boyer, L.A.: 1966,
Adv. in electro and electron phys. 22B, 885.
PROBLEMS OF OVERSAMPLING WITH SEC VIDICON TELEVISION SYSTEMS
E. J. Devinney
Introduction
1. Procedure
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 135-139. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
136 E. J. DEVINNEY ET AL.
q q
CHARGE CHARGE
DENSITY DENSITY
(a)
q
CHARGE
DENSITY
SAMPLE NUMBER
(e)
optical transfer function (OTF), for this process, let the first
measurement at Xo be written
00
M(xo) = 1
3
-00
Table 1
o o o
top where the contours have been moved down, and likewise for
the previous sample in the current line, on the left, where the
contours have been shifted to the right. These effects are more
pronounced in the small step case in Figure 2, right. The plus
(+) sign marks the address point of the beam, showing the shift
imposed on the point spread function. It is also clear that the
peak amplitude decreases. In the Fourier domain, it is clear
that there would be a significant imaginary component in the OTF.
For the one dimensional case we have the MTF curves of Figure 3.
The angular range of the phase transfer function (PTF) for the
selected step sizes is given at top. Also shown are the sample
frequencies for the same intervals. Tick marks on the curves
mark their equivalent width in frequency space. The arrow on
the 10 curve marks the frequency at which the PTF discontinuously
changes sign at 90 degrees. Relative to a non-destructive
Gaussian beam, we see that as the sample interval decreases, the
low frequency response drops, but the high frequency response is
greater than that for a non-destructive Gaussian. One sees that
there is some response even out near the sample frequency in
each case. Thus, the temptation arises to utilize a format
which crams as much onto the detector as possible. One could
reduce the scale of the image by a factor of two, for example,
and then step twice as fine to recover the same spatial inform-
ation. However, when one carves up the device into more photo-
meters, they rapidly lose their independence. In any real device,
reduced channel capacity will also occur; i.e., a reduction in
dynamic range sets in. It should also be noted that in a real
SEC camera, the size of the beam grows with distance from the
center, so that if one were working on the middle curve of
PROBLEMS OF OVERSAMPLING WITH SEC VIDICON TELEVISION SYSTEMS 139
5 zV2 5V2 5,
1.0
PTF(DEGREES)
0.9
zV2 (0,-'.01)
V2 (0,-42.80)
0.8
, (0,-90; 90,75.86)
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 141-155. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
142 A. BLECHA AND P. BARTHOLDI
It is evident that both photocathode and SEC target are not uniform. Further-
more, the efficiency of a photocathode will be colour -dependent, and the
target response will present some non-linear behaviour, especially near
saturation value. In general, it will be impossible to separate the photo-
cathode and target contribution to global gain, and so a calibrai:ion measure-
ment for each colour will be necessary in multicolour photometry.
4. DATA FORMAT.
The SEC camera used is scanned line by line, and a total of 600 lines per
picture is recorded analogicaly on magnetic tape. An off-line digitiser
produces a 1024 points record for each line. So the picture is divided into
1024 x 600 pixels ( Crane 1971).
5 • CALIBRATION.
Because of the spatial, colour and intensity dependency of the camera gain,
a series of calibration files is necessary for each colour to remove the
target non-linearity. It is made for "white colour". The camera is exposed
to uniform illumination (Crane 1973) during an increasing period of time
( from 30 to 360 seconds). Due to the good integration camera possibilities
( more than 8 hours ), it is equivalent to illumination intensity variation.
Further, two files for G and VI Geneva photometric system filters (Rufener
1971) are recorded in order to test the validity of the main calibration files
for coloured files interpretation. We thus have a signal exposure from 30
to 360 seconds at each point of the camera field. In most points, saturation
start gains importance only at 240 to 300 sec exposure, so we have enough
points to reconstruct at each point a calibration curve ( figure 5. 1 ).
Before initiating the calibration data handling, we have to examine data
compression gossibilities. Ie is not very convenient to handle continuously
15 files of 10 words each.
SOME EXPERIMENTS WITH AN SEC IMAGE TUBE 143
Figure 5.1
6. CORRELATION LENGTH.
Due to the rather arbitrary choice of element size (it was chosen so as
to optimise the readout SNR), some correlation between adjacent elements
may be introduced during readout of the tube. LetdN be the rms deviation
of data from the mean in a [ k x k ] square, andlN the mean of this rms
deviation in the many [ k x k ] squa~s through some fixed region of the
data. If there is no correlation, theiN value will be independent of k
and this value is about VN.
where N is the number of photoelectrons per
element. Following figure 6.1, we choose a reasonable size of elements.
C3 % N
Q.I
E
Q . I : size used for compression
a;
c 2
1Il
E
L- 1
c
d
Q.I
E
2 4 6 8 10 element size 20
Figure 6.1
144 A. BLECHA AND P. BARTHOLDI
no calibration 53.13
linear interpolation 28.67
quadratic interpolation 36.18
parabolic fit 24.86
There is also a physical motivation for our choice, in the sense that in
the expected dependency of the form
bJ ~ = 60sec ~ = 59,osee
DJ 'e= 30sec ~ = 25,55ec
91 ~,,"270sec ~; 268.Ssec
hJ ~=300sec ~'" 295 sec
Figure S.l
146 A. BLECHA AND P. BARTHOLD!
Figure 8.2
SOME EXPERIMENTS WITH AN SEC IMAGE TUBE 147
Thus we can use the fitted curves to linearise calibration files and examine
the efficiency of the procedure. The linearised one is almost filed with the
random noise patterns of one isolevel curve only. We used this test syste-
matically on most of the files (fig. 8.1 ). Starting with the 210 sec exposure
we can see local saturation on the upper edge of the picture.
Figure 8.3
We notice that some part of the picture at the boundary cannot be used
and that care must be taken to account for the local saturation effect wich
starts at the place of the highest camera gain. Figure 8.2 displays the
linear and quadratic coefficients of the parabolic approximation. Figure
8.3 shows G and VI coloured calibration files linearised by means of the
same coefficients extracted from the white light calibration files. The
linearisation is not completely succes sful, and so a special procedure
must be used for the calibration of coloured measurements.
9. ZERO LEVEL.
Zero level must be subtracted from all files ,treated ( calibration files
included ). The stability of this zero level used for subtraction was tested.
148 A. BLECHA AND P. BARTHOLDI
Level 20, 40, 60, correspond to the 2,4 ,and 6 sec exposure.
Figure 10.1
SOME EXPERIMENTS WITH AN SEC IMAGE TUBE 149
This test can serve also as general tube stability criterion if its variation
is within some background size limit. Unfortunately, we have to subtract
this level from the original points without the possibility of subtracting
some mean value, as it varies from point to point. This procedure also
increase the noise of the resulting file. Furthermore, two files of zero
level made before and after the calibration measurement are slightly
different (fig. 10.1). So we subtracted the first one from the calibration
files and used the second one as zero level for the measurement files.
S = S(x, y, I) 10.2
and so
f =1,.(1
cw
I' is an input flow calculated by means of the "white" calibration data,
and I is a known input calibration c coloured flow. As we are interested
in I regardless of any common multiplication factor, we can assume
1=1, and so f =1'. This means that we do not need to repeat all cali-
bration mea~Ji.ement for each colour, and that only one colour calibration
file supplies information about photocathode efficiency. In the absence of
coloured calibration files we can use the sky background level as constant
level calibration file. Unfortunately, as this value is generally only a few
times more than a noise rms value, this method can introduce important
errors.
150 A. BLECHA AND P. BARTHOLDI
Corrections
~
Programms for stellar ....... Intermediate
A
Magnitude calculation .... data
Numerical results
Figure 11.1
As the purpose of this work was to test the possibility of stellar photometry
with SEC vidicon tube, the astronomical data collected are those of the
galactic cluster NGC 581. As we have colour calibration files for VI and
G filters only, we used the sky background to correct the other colours.
A two-dimensional gaussian profile fit and line by line one-dimensional
SOME EXPERIMENTS WITH AN SEC IMAGE TUBE lSI
Figure 12.1
152 A. BLECHA AND P. BARTHOLDI
Figure 12.2
SOME EXPERIMENTS WITH AN SEC IMAGE TUBE 153
Figure 12.3
Figure 12.4
154 A. BLECHA AND P. BARTHOLD!
gaussian fit are compared with direct summing the flux above the background
Finally, the two-dimensional fit was selected as the best. A typical fit
of stars of magnitude 9 and 11 is shown in figure 12.1. The solid line
represent the calculated profile, while the calibrated data points are con-
nected by a dotted line. Figure 12.2 shows a typical display of the whole
calibrated field. Line by line display (a) is convenient for a control of
star recognition. On the other hand, an isolevel display (b) is more sensitive
to background gradient. Theoretically, a homogenous background should
be expected according to the calibration file in the corresponding colour.
Unfortunately, even in the case of the coloured calibra tion file available,
as for VI filter, a strong local gradient in the background is present. It
may be explained only by zero level uncertainty or data inconsistency.
Nevertheless, we carried out a series of photometric estimations of
stellar magnitudes in NGC 581. We compared them to the value of the
Geneva catalogue as given by Rufener (1971). (figures 12.3 and 12.4 )
If V and Bl measurements can be considered as compatible with the system
inherent noise ( essentially photon noise ), the U magnitude are comple-
tely beyond admissible limits. It can be pointed out that the large disper-
sion of experimental points is due rather to data incoherency than to some
lack of information about photocathode efficiency.
13. CONCLUSION.
14. ACKNOWLEDGMENT.
PROCESSING HARDWARE
FOUR YEARS EXPERIENCE IN APMS STAR PLATE PROCESSING - RESULTS
AND FUTURE PIANS~'c
James S. Newcomb
1. INTRODUCTION
In the previous session of this conference interesting and
significant advances in techniques of astronomical data acquisi-
tion with television, diode array, and electronographic
techniques with the possibility of data storage from these
sensors directly on computer compatible magnetic tape, core, or
disk storage have been presented. This paper is about a measure-
ment system for data that has been stored only on photographic
film - the measurement of stellar proper motions. More than
eight years of experience in digitizing the data stored on
photographic plates has developed in our group at Control Data
Corporation a deep and abiding respect for the vast amount of
information--wanted and unwanted, good and bad--that is stored
on each photographic star plate! The photographic plate, with
all of its non-linearity of response, granularity, and noise, is
still an outstandingly compact, practical data storage device.
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 159-170. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
160 JAMESS.NEWCOMB
~,\\
~
CDC 6600 Compu'"
Function·
-,
, .\
\ +~-
Function:
tempororycomputer
compotlbledoloslorog.
l"""'.
op'ical'phot~rophic
""
Function
imog.ono storpatl"ft
reconstruction,digilol
discovery ond measLlr,-
~ ~m:~~~
mentofsteltor prop.r
CATALOG Of PROPER
MOTIONS OF FAINT STARS .,
Function" '-
~~~~~~'~'e:f!~"'--
.......___..__
- F,,,lIon
electro-optical
~
~~
F,,,'ion'
bosicostronomicol
pholoqrophic interrogation r,f,renee data·
Figure 1.
FOUR YEARS EXPERIENCE IN APMS STAR PLATE PROCESSING 161
RECENT EPOCH
ORIGINAL STARPLATE
7
EARLIER EPOCH
DUPLICATE STAAPLATE
SCANNING SPOT
SCANNER REPORTS:
A Y IN MICRONS
SCANNER REPORTS
X POSITION
Y POSITION
SPOT POSITION
I -- SCANNER REPORTS:
OY IN MICRONS
1
"',
-
SCANNER REPORTS:
x POSITION
- - - Y POSITION
SPOT POSITION
SCAN PATH
",r·· '··'""[i
:J'F;:"~' .. ~ :.
l:?y.
)
.... :<1
~. ....... ~.
tt
. ..... .... -" ...
~(~ : : ~:
/ ····",,·j. \
--,:' . .~::~-\
,",om
CENTERS I' ...." . CENTER
CENTRAL TAANSIT
TABLE I
Candidate Motion Evaluation for Five Pairs of Star Plates
in the Range of 0'.'09 to 0~'18 Annual Motion
Number of
Number of Percent
Quality Motions
Evaluation Motions of Total
Index with this
Rejected Rejected
Index Range
Acceptable Q < 24 1514 27 1. 78%
Doubtful 24 < Q < 30 527 235 44.5%
Unacceptable Q > 30 62 50 80.6"/0
TABLE II
Listings of Proper Motions in the Series Entitled
"Proper Motion Survey with the 48-Inch Schmidt Telescope"
Series Date
Title
Number
XXVIII Faint Proper Motion Stars near the December, 1971
South Galactic Pole
XXX Proper Motions for 1357 Faint Stars March, 1972
XXXI Proper Motions for 2520 Faint Stars June, 1972
XXXII Faint Proper Motion Stars near the September, 1972
South Galactic Pole
XXXIII Proper Motions for 3478 Faint Stars November, 1972
XXXIV Proper Motions for 4452 Faint Stars July, 1973
XXXVI Proper Motions for 6955 Faint Stars February, 1974
XXXVII Proper Motions for 4483 Faint Stars July, 1974
7. CONCLUSIONS
The measuring system developed for the measurement of proper
motions from the plates taken at the 48 inch Schmidt telescope
has shown that the automation of the discovery and measurement
of proper motion can speed up the documentation of a statisti-
cally significant body of astronomical data. The time has come
to offer to the rest of the astronomical community the opportunity
to discover, measure, and document other significant bodies of
astronomical data; Control Data Corporation welcomes this
opportunity to serve astronomy, and looks forward to meeting the
challenges of their data measurement and reduction problems.
REFERENCES
1. LaBonte, A. E., "Automated Proper Survey: Data Reduction
and Initial System Performance," Proceedings of the
International Astronomical Union Colloquium No.7,
University of Minnesota, Minneapolis, MN, December, 1970,
pp. 26-47.
1• INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 171·184. All Rights Rese11Jed.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht·Holland.
172 C. A. MURRAY AND W. NICHOLSON
2. SEARCH PROCEDURE
x 30 1024 1
X 75 409.6 0.4
X240 128 0.125
I, = :~ Teo s 9 d 9 dr = 0
~ r:;
R, 0
Il R'rR, 0
TGin ed9d r = 0 ( 1)
I3 - -Jr w R'r
R,
()M
0
(T-P) d" dr = 0
P
clr aM
2'" ?/ r
M ~ r
>/
~
M
M-C
1
-.
1 + (r-M)C
0
-1
C
0
-c
-i
M-C ~ r » 0 0 0 0
The integrands in (1) are zero for r>M and r<M-C. Further,
since the star profile is only approximately linear over part of
the radius, the range of integration is reduced by two parameters
Ill' Il~ to the range M-Il, >"r~M-Il,2.. for M>,Il:l + R,1 (saturated images)
and M-Il, ~r~R:t for M< Il,. + R;&.. Provision is made for changing
the profile function at M '" Il,1 + Ra to one with constant radius
and decreasing slope, as described by Walker (1971) but this
system has not been used at R.GoO.
M, ~ 2 '"
-If-
2 - 1
D, ~
2'.:1 _ 1
D,. ~ 2'" - 1
S ~ 2" - 1
where S is the slope parameter which is related to C by
;1.0-1\
SC =0 2
1: t
The image profile T and the cumulative integral
111
r~, , =: T - P') d6 dr
M-D, ()
are displayed on an oscilloscope, as sho~~ in Fig.1. Before
starting measurement, the operator selects the profile parameters
empirically from examination of the shape of I~~r on typical
images, the aim being to set it as straight as possible.
(a)
(b)
Fig.1
(a) Transmission T
-------------02 - - - - - - - - - - - - -
o - - - - - - - - - - C ---------------
R,.2 1O
Fig.2
178 c. A. MURRAY AND W. NICHOLSON
Fig.3
Data from all plates measured on a field are accumulated on
to magnetic tape by successive use of Programme 3. The general
philosophy underlying the merging procedure is similar to that
used in matching images in Programme 2, except that matching is
performed after independent 3-constant linear transformations have
been applied to each coordinate, since the plates may well have
been taken on different telescopes under a "ride variety of circum-
stances.
6.1. Coordinates
"
... ... II ..
... ... ... ... I."
". ,
•
•
··u 11"
,··················1I .. ,U'U.. ·,·····_·· •••••••• ..1011
au.. '"
I"
................................_ ............... .
• • UII'
o .: •••••••••••••• :;i:l~U1:i.~ ................ :.:.:~.: .......;........................................
oJ.. 'u,'. ,": I I
(a)
::::~ (b)
M 'N
•
·
• ;0111 U II ..
• 1 II
• III I "
"" I
II
n
I II
lIn
Fig.4
o
liM
-10
...
.....- ......
-
o
0
0
0 o
. ....•
••
0
• o
o
o .0 0 o • o o o
• •• o • _• 0 o
o
-20 o 00 o o
(o) o o
14 13 12 Magnitude 11 10
+5
t.M
o
-5
Ib)
Fig.5
THE USE OF THE 'GALAXY' MACHINE AT THE ROYAL GREENWICH OBSERVATORY 183
6.2. Magnitudes.
7• FUTURE ENHANCEMENT.
There have been t.'o main causes of lost output in the system
during the past tv'o years, failure of the paper tape input/output
equipment, which accounted for about 75 per cent of "doyn" time,
and unacceptable drifts in the datum, referred to in section 5.
Both of these should be considerably reduced by the introduction
of a mini-computer into the system.
8. ACKNOv-JLEDGEMENTS
Our thanks are also due to Mr. Khao Muanwong and Mr. P.M.
Corben for useful discussions, to Mrs. M. Everest and Mr. D.
Calvert for preparation of diagrams, and to Mrs. E. Howard for
preparing the "camera-ready" typescript of this paper.
REFERENCE
.
J.R.W. Helntze0) ,R.L. Porteous00) and W. Brandie00)
I. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 185-198. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
186 J. R. W. HEINTZE ET AL.
1cm
x - carriage trace 1 _ _ _ - - - - - - - - ...... __ _
,.m [~::':~~~:~~~~~~~~:~m~~:::~0~::::::::cc~;~1
traceZ-- - - -
optical flat set parallel to
carnage motion
situation
:~;r~i~~:ce~ /t ~"'. and duri~~L._t_r_ac-c_e--,2=--+--,-_~
" _ x carriage
~I_ _-+-~
stationary
micro-comparator
probe
20 r -__________. -__________~----__~~~~~~-------r----------_,100
---' -- ""-.
..
TRANSMITTANCE OF
I>/~ ., DICHROIC BEAMSPLITTER
SPLITTING VISUALJAUTO-
/ Q.E OF 5-11
...... , .
FOCUS BANDS;
./ PHOTOCATHODE
'\
....
\
15
\ .\ 75
,..
;;-
".g
zu
w
\ . z
w lL \ oiJj
u lL
Zw '"~
~~ 5O~
u=>
wI- «
--'Z cr:
lL:,\ I-
IIiiI1:1~ I I
!l!o
I ,I
II
5 REFLECTANCE OF
-!.
DOUBLE COATED SURFACE
".
Jr __
Jr
300 400
I - - - ---,,~,::::.:
I
I
I
I ~
I !"
I ;:;:
I SUB PLATE =:
I OPTICS t!:I
.... focus channel A.:.. 650 nm I
I ~'"'''e I ~
_____ vlewlng channel A.= 560 nm N
_ _ measurement channel A.: 485 nm L_______________ J t!:I
t!:I
>-3
Figure 5. Schematic arrangement of optical channels. ;J;>
t""
light pipe to autofocus unit
-- --- -- - -- -- - - -- ----- ------- - - - --,
=0
~
'rl
c:::
r
arJ"l'.Jar wedge SLIT PROJECTOR ~
(adjustable} :>
~ i ~
I o
t is:
--...::.-.0.:::.-=----
t :>
·aperture stop t
compensat ing wedge slit length adjustment 0-18 mm
::l
(")
o
=0 ~Sllt width adjustment 0-1.5 mm; .OOSmm dlv. "C
~== :>:J
L--------------------------l o
t --..... rot~tion unit C'l
r------~v>-- - -- - --- - --- -- - r--::::--' I 90 ; Ofdlv. :>:J
:>
/1 I 1'1\ , - - --- - - ffir/fixed ~klgr.> : el :: I is:
K \/\r . :: ::
"II 1\1
II n
o
I I
----l t ,~I'I z
I 0 I I I >-l
''01 :>:J
1"5 II II I :I : tube 'ens
I 0 , or
L..J1J_J I 1/ narrowband
r
interference fitter m
tl
is:
n:>:J
o
tl
m
,-----'-- Z
t
t
'"
::J
I o
I is:
I
m
I t;l
t
t ~
t o
I is:
"C
:>
SUB-PLATE :>:J
OPTICS
~
o
:>:J
3. OPTICAL SYSTEM
- - - - - - - ftir.-.fr-Ilr""Ikllr-
- - ~ -
4. CONTROL SYSTEM
0.9,-,-,-,-"-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,-,,
0.8
0.7
o
I I I I
o :1·9-·-·-------·-·-----------·------------- c-11
, ,
o~ ________________~:[ i
:r-----------------~
, ,
:A D:
+--< ----- --- --------- ------------------IJ
~e~h~n!c~l_sZs!e~
Working area 251 cm x 103 em
Measured deviation from orthogonality of
mean carriage motions 1.1 + 1.2 arc seconds
THE FULLY AUTOMATIC PROGRAM-CONTROLLED MICRODENSITOMETER-COMPARATOR 197
~e£h~nic~l_sys~ (continued)
Absolute accuracy of the gratings. on each
axis at 20 0 C + 0.8 ~m
Minimum step size 0.5 ~m
Repeatability of position measurement + 0.35 ~m
Operating speeds:
fast manual traverse 11 mm/s
slow manual traverse 0.25 mm/s
"on the fly" under computer control 0.1 - 3 mm/s
times for stepping, settling and density
measurement at D = 3 for 5 ~m steps 39 ms
25 ~m steps 40 ms
Qp,!.i.£.a.l ~y~t~m
Minimum slit widths and maximum slit lengths
25 x objective 1 ~m; 0.78 mm
10 x objective 2.5 ~m; 1. 95 mm
4 x objective 6.5 ~m; 4.90 mm
Range of focussing movement 3.25 mm
Focussing resolution 0.45 ~m
Orthogonality of focus motion to spectrogram
plane within 1. 75 minutes of arc
Acceptable slope of spectrogram 1 :80 at 3 mm/s
Focus error band of
25 x objective
-+ 1. 35 ~m
10 x objective !. 4 ~m
4 x objective + 19 ~m
REFERENCES
Edwin W. Dennison
MECHANICAL DESIGNS
de Jager/Nieuwenhuijzen (eds.), /rruzge Processing Techniques in Astronomy. 199-208. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
200 EDWIN W. DENNISON
OPTICAL CONFIGURATION
SETTING DEVICES
(a)
(b)
(c)
CONPUTER SYSTEN
~
~
tv
o
-J
208 EDWIN W. DENNISON
SUMMARY
REFERENCE
K. Aa. Strand
U. S. Naval Observatory
Washington, D. C.
ABSTRACT
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 209. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
THE TWO-COORDINATE GRANT MACHINE IN LONG-FOCUS ASTROMETRY
John L. Hershey
1. INTRODUCTION
de Jag:r!Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 211-216. All Rights Reserved.
Copynght © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
212 JOHN L. HERSHEY
3. PERFORMANCE
The 25,000 plates measured on the Grant machine during the past
few years yield many comparisons with measurements of plate series
by visual bisection with cross wires on the old "classical" one
screw machines. Usually a decrease of 30% in the error of reduced
star positions is achieved with the Grant machine, corresponding
to a doubling of plate weight. This accuracy is achieved with one
setting per reference star image of the Grant profiles as compared
with four settings of visual bisection with cross-wires. The high
214 JOHN L. HERSHEY
Most of the plates haye been measured with only one setting
on reference star images and two settings on the central star
,.;hich opens and closes the measuring sequence. The typical error
of repeatability of the position of one image after plate reduction
in a stellar reference frame is approximately one micron which is
less than the intrinsic error of the photographic process. Exper-
iments with several plate series have been made with two indepen-
dent measurements of each plate. By comparing the reduced positions
from two measurements on each of several hundred plates the error
of one measurement of the Grant machine has been found to be ap-
proximately one-half the error of the photographic positional
information of images in a stellar reference frame. Thus the
maximum improvement possible in the present photographic technique
by a perfect measuring machine should be only about 10%. This
limit would predict a five percent decrease in probable error when
the mean of two independent measurements is taken. Reductions in
error of 5% or less have been found experimentally where parallax
solutions have been made on several plate series from single meas-
urements and from the mean of two measurements. It is a matter of
personal opinion whether it is wiser to double plate processing
time or increase the number of plates by 10% or less to achieve
an equivalent level of accuracy.
4. CONCLUSION
REFERENCES
Lippincott, S. L., and Hershey, J. L., 1972. A.J. 77, 679. "Orbit ,
Mass Ratio, and Parallax of the Visual Binary Ross 614."
van de Kamp, P., 1973. A.J. 78, 1009. "Parallax and mass ratio
of the visual binary 61 Cygni."
THE COSMOS FACILITY AT THE ROYAL OBSERVATORY EDINBURGH
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 217-222. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
218 N. M. PRATT ET AL.
The operation of this mode makes use of a linear scan CRT together
with continuous controlled motion of the Y-Axis. The position of the
spot in the scan is monitored, permitting this to be directly related to
the main carriage measurement system. The optical components image
the scan in the plane of the emulsion with a magnification of i. The
scan operates on the basis of a fixed number of increments - 128 - whose
size may be chosen to be 8,.u.m, 16;um, or 32}J-m, the spot size being
adjusted to match the chosen interval. The plate is scanned in lanes of
width 1 mm, 2 mm or 4 mm respectively.
Hardware logic and storage analyse the images and produce 8 para-
meters/image with position information at the end of each scan in which
one or more images terminate. These are transferred to magnetic tape.
The limiting transfer rate is 4,000 images/sec which gives a handsome
margin on the actual average image density of less than 20/square mm,
giving 40 images/sec in practice with an 8,;'-Am increment.
220 N. M. PRATT ET AL.
(a)
~~
5
~ 3
cd
u
(b) u:l
.....
0
2
......
~
cd
......
u:l
1
E Y(Xmin
E
+ Xmax)
change
}for calculation of centroid
2. MAPPING MODE
The output of the photomultipliers is integrated over the increment
distance using the same scanning methods as for coarse measurement.
The transmission value is digitised to an accuracy of 1% and transferred
to magnetic tape.
This mode allows the detailed examination by off line computer
analysis of areas of doubt or complexity such as bridges between galaxies
or galaxies with complex structure.
3. FINE MODE
Having determined the positions and other parameters of images
either directly using the CM or indirectly via a computer analysis of a
mapping mode output, a punched paper tape containing the plate
co-ordinates of interesting images is prepared at the off line computer.
This tape is used as input to the fine mode and will contain the co -ordi-
nates, ellipticity and an approximate orientation for non-circular images.
The fine mode utilises the alternative densitometer system which con-
sists of a generalised elliptical scan generator driving a micro focus
CR T together with the optics necessary to focus an image in the plane
of the emulsion. The generator produces a concentric family of 1024
ellipses sequentially. In this mode a large magnification factor may
be employed 1;240, which entails the use of a 4 mm flat field objective,
the depth of focus of such an objective is small, very.much smaller
indeed than the deviations from the plane of the plate itself. In order
to overcome this, the objective rides on the surface of the plate, the
separation being provided by an annular collar, with the force on the
plate being reduced to 3 gm by a pneumatic balancing system.
222 N. M. PRATT ET AL.
REFERENCE
1. Dodd et al., A study of 3000 faint Galaxies, M. N. R. A. S. (in press).
THE ESO S-3000 MEASURING MACHINE
R. M. WEST
1. INTRODUCTION
For those involved in the ESO Schmidt project, the need arose
early for some way of evaluating the images on these plates. Dis-
cussions were undertaken among astronomers inside and outside ESO
and as a result it was proposed to acquire a measuring machine
that would be capable of measuring large Schmidt plates at a reaso-
nably fast rate. Main restraints for the choice of machine were of
budgetary nature. The following tentative specifications were set :
1) measuring area 35 x 35 cm, 2) positions to + 1 ~m, 3) data rate
at least 2 kHz, preferably 5 kHz, 4) density t~ + 0.02 D in the
interval 0-4 D. -
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 223-228. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
224 R. M. WEST
2. HARDWARE
The S-3000 was designed for a 5 kHz data transfer rate, but tests
show that the present logarithmic amplifier is not quite as fast,
and we have therefore prefered to use a 2 kHz rate for the moment.
A new and faster circuit is being developed by the ESO-TP Division.
3. SOFTWARE
The programme collects the positions, the radii and the total in-
tegrated density over the image surface in a disk file. This file
is then be used for further reduction, as for instance, when astro-
metric and photometric standard stars are available in the field,
to compute equatorial coordinates and stellar magnitudes.
THEES()'3000 MEASURING MACHINE 227
The positional r.m.s. accuracy, which has been estimated from re-
peated scans, is about ~ 0.8 flm in X and + 1.1 flm in Y for the
geometric center. The r.m.s. of stellar r~dii is ~ 0.5 flm for
images with radii between 20 ~m and 80 ~m, the medium size images
being the most accurate. These figures correspond to 2 flm steps in
X and 10 ~m steps in Y. When repeating a scan after 4-6 hours,
small zero-point offsets are noted (about ~ 1 ~m both in X and Y).
4. SUMMARY
The experience with the ESO S-3000 measuring machine has shown
that it is capable of performing useful astronomical measurements
of Schmidt plates at a reasonably fast rate by means of two-dimen-
sional raster scanning. The application of a diode array would
shorten the measuring time with a factor equal to the number of
diodes. The installation of a diode array in the present machine
(substituting the slit) would therefore greatly enhance its capa-
city. Studies are now under way at ESO to investigate the feasabi-
lity of such a system with on-line reduction.
ACKNOWLEDGEMENTS
REFERENCE
1. R.M. West, ESO Bull., 10, p. 25 (1974)
IRIS - A TWO AXIS CO~PARATOR AND HICRODENSITO~,1ETER lJSING TWO
DIFFERENT SCANNING ~,nDES
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 229-236. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
230 N. ASLUND ET AL.
INTRODUCTION
An early example of the use of two different scanning systems in
a measuring machine is given by the so called Spiral Reader. This
machine was originally built by prof. L. Alvarez at Berkeley to
measure bubble chamber plates. Here a fast rotating system pro-
vides for an efficient use of the measuring time by allowing a
de~red concentration on the objects of interest. The movement
from one such object to another is accomplished by moving the
measuring tables.
The same basic philosophy has been adopted in IRIS. However, IRIS
is a general purpose machine, that can be used to measure any
kind of patterns recorded on photographic plates. Hence the fast
scanning has been designed to supply very basic information about
the density distribution around the measuring point, e.g. the
size and direction of the density gradient. Like the Spiral Reader
this new machine has been designed to work with a computer.
Further, the idea of making use of interactive procedures has been
extended and has led to a design, that allows efficient communi-
cation between three parties: the operator, the measuring machine
and the computer.
These objectives sometimes necessitate an unorthodox desi~l, since
they may lead to conflicting demands. As an example we may take
the problem of illuminating the plate. To display the plate to the
operator, a large part of it has to be illuminated. On the other
hand the photometric readings communicated to the computer should
not be distorted due to straylight. This raises a demand for a
very small illuminated area. The method chosen in IRIS is to use
light of different spectral composition for the different purposes.
Another example comes from the demand that the operator and the
machine be close to each other, physically, which is in conflict
with the demand that mechanical or thermal disturbances should
be avoided.
In addition there have also been some more general principles
employed in the design work:
1) To acquire the high accuracy (both in positioning and photo-
metry) rather by automatic calibrations than by large-scale
IRIS - A TWO AXIS COMPARATOR AND MICRODENSITOMETER 231
f?'I ..... .
.'
. ....
'
Fig.3. Reference channels for positioning (in the front) and for
intensity (fiber optics) allow continous comparisons with
fixed references. The plate is indicated in the figure but
not the measuring tables.
234 N. AS LUND ET AL.
measurements.
This is also done when testing the machine as a comparator. A
precision grid plate is measured, and the machine makes automatic
settings on the crossing points of this grid. This is possible
since the machine has the ability of setting on the median in the
x- and Y- direction of any intensity distribution on the plate,
Aslund (1965). The (rectangular) areas of integration employed at
these settings can be choosen at will. In the present application
the median setting is used both to make the machine follow the
mesh lines and to make it perform repeated settings on the cros-
sing points.
During these measurements, which allow a separation of the random
and the systematic errors, the computer also collects data from
different sensors placed at arbitrary points of the machine, mea-
suring voltages or temperatures. A data analysis is performed to
reveal any correlation between the systematic errors observed at
the measurements on the grid and the changes of these parameters.
In this way it has been possible to trace the main error sources
in the machine and to take appropriate counter measures. Presently
the environment of the machine is not temperature controlled but
at night test, with temperature changes within ± 0.5 0 C, the machine
has remained stable for several hours within ± 0.5 micrometers.
REFERENCES
As lund , N: 1965, Dissert., Dept. of Physics, University of
Stockholm.
MACHINE A MESURER DE L'OBSERVATOIRE DE HAUTE
PROVENCE
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 237-243. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
238 CH. FEHRENBACH AND A. YIN
Figure 1.
AX = AX + B
and Xli = Xi - AX
SPECTRE DE COMPARAISON
A = +.9501407E-04 B = -.1046769E+02
CONTROLF. OPERATF.UR
-211
SPECTRE DE COMPARAISON
A =
+.1125105E-03 B = - • I 19 Ii:-ll 3 F.+ 0 2
RAIE ABSCI SSE iHEORI (,1llE DI FI'l':>1F:MCE RF:SIDU RAIE
05 +()1'>3699 I A - 0 614 -01 1 7 ()5
06 +()1'>7n?4I R -04 1 3 +00 1 1 06
07 +072471'>1? -0414 -00 1 1'> 07
OR +0762671 R -02 1 4 +00 1 9 ()R
09 +07020K 1 7 -04,:l -()113 ()9
10 +()R~52112 -02 1 3 +00,3 10
11 +()RR4521 /j -0 110 +00 1 4 11
12 +o043RI'>,7 +0113 +02 1 1'> 12
13 +009 0 09, I~ +00 1 1 +()(),8 13
14 +lfJl7KO l ll -00 1 7 -()O,2 14
15 + I OA21 5111 +0012 +()(),2 15
16 +113531,0 +0 II 6 +0017 11'>
17 +IIR976,3 +00 1 7 -00,8 17
IR +121730,5 +0 1,3 -00,5 18
19 +127737,5 -00, I -02,6 19
+ 1361 59,8 +0 .. ,3 +00,9 20
20
CONiP()LF. OPERATEUR
Figure 2.
MACHINE A MESURER DE L'OBSERVATOIRE DE HAUTE PROVENCE 241
I: ( AV)2
0'=
n ( n - 1 )
CONTROLE OPERATEUR
-O~,
SPECTRE ETOILE
VR :: +021.0 KM/S
RAIE AsseI SSE THEORT QUE VITESSE RESIDU RAIE
01 +072478,fl +019.2 +0 I. 7 01
02 +n7696n,1 +021.9 -()),I) O?
03 +081596,3 +020.5 +00.4 03
04 +062S?I.O +016.0 +02.9 04
05 '+083969,2 +020.4 +()O.5 OS
06 +084630,8 +021. I -00,2 06
07 +f)~7082,5 +022.1 -01,2 07
09 +09'139 I,~ +020.1 +00.6 n9
10 +0968QI.6 +02~~. I -02.1 10
II +099157.£1 +020.1 +00.8 II
12 +IOI161,n +026.8 -05 .. 9 12
13 +105301,7 +O·,,~.7 -01.8 13
14 +106493,7 +O2:!.2 -02.3 14
15 +11I7C.)7,3 +019.5 +01.4 15
16 + 116866.4 +021.8 -00.9 16
17 + 1 174 7fj. 9 +U!8.2 +02.7 17
18 +118973.2 +015.9 +O~,O 18
19 +121 7 3<::.7 +U20,1 +00.8 19
20 +1l2;Jil27,7 +C:2().5 +L\U.L! 20
21 +123786.7 +()21.2 -00.2 21
S1 GMA : +0.45
CONTROLE OPERATEUR
Figure 3.
MACHINE A MESURER DE L'OBSERVATOIRE DE HAUTE PROVENCE 243
I JPlCTRt I
"--- t.CROU
HI\~UR =-~ ----- f - - - - - - --e-3-+--
VIS
I r l Rf.G.Lf.
I
r
CO,M.UR
I
t
ORblHAT{.UR HULT1.8 r----- PUPITRf.
( ln~techniqve I
16 J( motJ tJ bitJ
TTY
AUTIJHATIX _ IJ..J.I.P
Figure 4. -
REFERENCES
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 245·246. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht·Holland.
246 E. J. KIBBLEWHITE ET AL.
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 247-253. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
248 P. W. H. BLANSJAAR AND J. VAN KUILENBURG
Spotsize min. 10 fm
Accuracy I fm
Stepsize 5.0 fm
Dynamic range I : 100
l1ax. density 3D
Density read-out 10 bits
Diode array 128 elements
Opt. read-out time: 3msec (128 el.)
Hax. speed 4 mm/sec
One-step speed 20 msec/step
Hono Comparator David W. Hann
Computer PDP 11/40
Hemory size 32 Kbytes
Control language Fortran IV
THE COMPUTER CONTROLLED COMPARATOR ASTROSCAN 249
ASTROSCAN
interface Comparator/Densitometer
Com uter
Diode-array
PDPII/40 ~ ______~ Controller Optic,s
Image
X-y motors
Lightsource
°,°,
DO 2 1= 1,10 step in Y direction
CALL HOVE (1 1) run forwards in X for backlash
DO 3 J=1 ,200 step in X direction
CALL DENS optical read-out
WRITE (5.4)(ID(I).I 1,128 print result
4 FORl1AT (' ',2014)
CALL HOVE (1,0, 1 ) I mOve In X direction
3 CONTINUE _J
CALL HOVE (-210,256,2) move In Y direction and backwards
2CONTINUE
CALL EPIL
STOP
END end of program
REFERENCES
de Jager/Nieuwenhuijzen (eds.), 1I1Ulge Processing Techniques in Astronomy. 255-260. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
256 D. H. ANDREWS AND K. O. WRIGHT
MONITOR
"
GQ) ~
H N
EMI 6094
="
~PAE5LrT
SM~---Ee=~-
~ HOLDDOWN STRIPS
A
PLATE MOTION TO MEASURE
CAllBRATtoN SPECTRUM
=L4
VIEWING LAMP
WOBBLE PLATE
RLS
ANALYZING SLIT
o(PR=:RY()
V LS EMI 8094
VR SIGX
GI P31
TAPE
NOVA UNIT
1220
451.P'S.
VR
~
CHART RECORDER
En = AT -~
Eg = 0.074(1-T)
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 261-264. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
262 ROBERT J. RUTTEN AND H. JACQUES VAN AMERONGEN
ACKNOWLEDGEMENT
REFERENCES
Brault, J.W. and White, O.R.: 1971, Astron. Astrophys. l}, 16.
Gratton, L. et al.: 1971, Pub. Roy. Obs. Edinburgh ~ nr~-II, p. 142.
Hoekstra, R.: 1969, Thesis, Amsterdam. -
PHOTOELECTRIC SPECTROSCOPIC PLATE MEASURING MACHINE BUILT AT
THE MEUDON OBSERVATORY
Frangoise LAUNAY
INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 265-268. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
266 FRAN90ISE LAUNA Y
DESCRIPTION
The figure shows the layout of the optical system. The plate
is horizontal, however the plate-holder has been carefully
designed to avoid any bending of the plate under measurement. Of
course the plate-holder can be rotated in order to have the
spectrogram normal to the travel.
field diaphragms
plate
viewing
screen
~;;~~:J~:t==========::~::::~~~d2 dichroic
plates
rotating
PM] PM3
REFERENCES
SOFTWARE TECHNIQUES
SOFTWARE TECHNIQUES FOR IMAGE PROCESSING
Daniel A. Klinglesmith
1. Introduction
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 271-283. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
272 DANIEL A. KLINGLESMITH
r---~
Input
Simple
Language
I
I Language
Translator
I
I I
Input
Images -I I Task
Scheduler
I
I Applications
Program
Library
I I
Output
Images -L 1/0 I
r+-+ Routines
I I
Non-image I lCurrent
Output I--
I Applications
L ___ ~I
Program
Table 1
Victor E. Shely
U. S. Army Engineer Topographic Laboratories
Fort Belvoir, Virginia
H
H
o...H
en enenHHenenen~o...
1J.l~::;:uu OOo...<t:H
::;:o...<t:IJ.lj:l..XUUHUU
Image Combination HenOO(f.)<t:IJ.lIJ.l::;:H....:i
OIJ.lHH::;:o...o...o...en>;;;::
Mosiacing X X X XXXX
Differencing X XXXX
Averaging XX X XXX
Linear Combination, General X XX X
Ratio X X XXXX
Histogram Equalization X X XX
Correlation X X X XX
Modification X XXX
Image Enhancement
Contrast Stretch X X X XXXXX
Image Shading XXX X XX
Convolution Filtering X X X XX
Recursive Filtering X
Non-Linear Smoothing XX XX
Geomtric Manipulation
Change of scale Linear X X X XXXX
Change of scale Non-Linear X X X XX
Rotation, General X XX
Rotation, 90° only X XXXX
Transposition X XX X
Frequency Domain Analysis
FFT XXX X XXXXX
Filtering XXX X XXXXX
FHT, (Hadamard) X X
Filter Generation X X X XXXX
Utilities
Image Generation X X X X XX
Grid Overlay X X X XXXX
Printer Listing XXX XXXXXXX
Image Copy XXXXXXXXXXX
Histogram XXXXXXXXXXX
Text Generation X X X XXX X
276 DANIEL A. KLINGLESMITH
• • •
4. Conclusions
References
H. Nieuwenhuijzen
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 285-299. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
286 H. NIEUWENHUIJZEN
"Filter"
Deconvolute
t
Instrumental
Telescope, data acquisition profile
I
control (CAHAC)
processed
data
----.-o!i..
FEED IN ....
FEED OUT
========
PRESENTATION
============
A: TEID1S IN WHICH TO DESCRIBE
B: SOFTWARE TO DO THE PROCESS
under terms described
C: IN REALITY: 2-DIH
INFORHATION Figure 1
PROBLEMS OF MACIDNE INDEPENDENCY AND INTERACTIVE MATTERS 287
preliminary copy of this list for immediate availability, to give
comments and make any changes that seem desirable. The list has
been revised before final publication.
2. RESPLTS
A. Interactive processing
B. Classification
From the material presented it seems that there are three main
areas in which consolidation of software effort has taken place.
i) The picture processing systems have a large amount of software
modules available for general methods in picture processing.
Many of the programs have been originally developed by the
Jet Propulsion Laboratory for Apollo and Mariner type photo's,
and have been used and augmented by various institutes. Those
who are interested are requested to contact Drs. Dunford and
Klinglesmith.
PROBLEMS OF MACHINE INDEPENDENCY AND INTERACTIVE MATTERS 289
ii) There is some consolidation in the recognition cycli for
stellar and nebular images. It should not be forgotten that
for example the Luyten measuring machine (see the paper by
J.S. Newcomb in the proceedings of this conference) has
buried in it a large amount of software, as has also the
Galaxy machine.
iii) Also there is quite a degree of consolidation in the field of
one-dimensional image processing.
See also the classical paper by Brault and Hhite (1971).
3. LIST OF PROGRAMS
3. I. a) Image combination
b) Image enhancement
c) Geometric manipulation
d) Frequency domain analysis
e) utility programs I/O
N
Institute Contact person Software Computer Core \0
Remarks o
Laboratory for Atmospheric Bruner, E. OSO-I
and Space Physics, image system,
University of Colorado, interactive picture
Boulder, Colo 80302, USA. processing
2-dimensional
I-dimensional
Science Research Council Dunford, E. Image processing IBM 370/195 modules VICAR
Appleton Laboratory software for IUE - 10 k program
Slough, United Kingdom satellite, developed - 50 k
from Jet Propulsion -300 k
Laboratory for pro-
cessing Apollo and
Mariner Pictures
See also Klinglesmith
NASA/Goddard Space Flight Klinglesmith, D.A. IBM 360/75
Center, Maryland, USA. IBM 360/91
Center for Astrophysics Levine, R.H. Smoothing, filtering CDC 6400, Harvard ex-
Cambridge, Mass. 02183 scaling, image pro- Comtal 8300 periment on
USA. cessing, interactive display + Skylab Mount
display PDP 11/10 (20.000
spectrohe-
liograms to
;:r1
be process- z
ed) at up t:a
to 7 UV ~
tTl
wavelengths Z
::I:
simultan- c::
eous t::
N
tTl
Z
Institute Contact person Software Computer Core Remarks ;g
ot:!j
American Science and (R.H. Levine) Analysis of solar f;;
Engineering X-ray photographs
:s:
'"
o
Cambridge, Mass. 02183 'Tj
USA. :s:
:>
n
Johnson Space Center Lyon, J. On-line smoothing, IBM 360/75 Earth re- 23
Houston, Texas, USA. in liaison for expanding, + on-line sources ~
solar physicists pattern recognition display Information 52
hardware Processing ?il
;:g
(ERIPS) z
system, used ?il
' ,
b y SC1.ent1.sts z
n
k '
wor"1.ng ' h
W1.t - <
:>
Skylab data ~
from the 52
Harvard expo t;J ::<:l
MIT Image Processing McCord, T.B. Interactive and IBM 360/65 158 k batch system q
Program, Kinnucan, P.F. batch processing for 4 years,
MIT, USA. Fawcett, G. In. 2-dimensional images interactive
~
:s:
:>
and for spectra, 1 year, >-3
>-3
many programs. operative tr1
::<:l
programs developed in '"
interaction with JPL,
Stanford U. Research
Inst. and US Geolo-
gical Survey
Kitt Peak National D. Wells Interactive CDC 6400, See article
Observatory, data processing Varian V 74 by Living-
P.O. Box 26732, Tucson + Comtal 8300 ston
tv
Arizona 85726. USA. interactive \0
display
3.1. f) Reduction of radio maps N
1.0
h) Interactive processing of radio maps N
Sterrewacht Leiden Van Someren Greve, IBM 370/158 192-256 k PLI language
Ruygens Laboratorium R.N. IBM 1800 interactive
Leiden, the Netherlands
Sterrenkundig Laboratorium Allen, R.J. interactive pro- PDP-9 24 k Assembler
"Kapteyn", Roogbouw WSN, cessing
+ 10 6 Fortran
Postbus 800, Groningen
the Netherlands
Astronomy Group Stephens, C.L. examines star field, CDC 6400 12-24 k Fortran
Imperial College fits curves, CDC 6600
Prince Consort Rd displays contours,
London SW 7 2BZ shows cross section
England at angle 8 to x,y interactive
Dunsink Observatory Wayman, P.A. Galaxy photometry Data 2 x 16 k New:
Dublin, Ireland General + 10 6 BASIC
ESO-TP Division West, R.M. recognition and RP 2100 12-32 k image
c/o CERN, measurement of classific-
CR - 1211 Geneva 23 stellar and extended ation
Switzerland images (Compare
with Luyten
Fortran)
iterative c:::
methods ::;::
t!j
smoothing z
and Fourier ::c
S
.....
processing N
t!j
Z
"C
3.2. c) On-line with computer to microdensitometer or flying spot scan
ottl
'"
r
Institute Contact person Software Computer Core Remarks M
::;::
NIWARS, Delft Van Kuilenburg, J. star search and PDP 11/40 16 k Fortran o'"
'"!j
Royal Greenwich Observa Murray, C.A. "Galaxy" star search; Nova 2/10 changing ~
1);
vatory, Hers tmonceux , Nicholson, W. sorting and collating over to
United Kingdom data from different on-line
plates, reduction facility
tV
\0
VI
N
3.3. a) One-dimensional time series \0
0\
Kitt Peak National Brault, J. extremely powerful CDC 6400 50 k multiple pass
Observatory, Slaughter, C. general reduction
P.O. Box 26732, Tucson, et al. program: Reducer
Arizona 85726, USA.
Laboratoire d'Astro- Delcroix, A. reduction of micro- Burroughs 13 k Algol
physique, Mons, Belgium photometric data, B - 6700
calibration
Kitt Peak National Furenlid, I. reduction of spectra CDC 6400 32 k multiple pass
Observatory, including image tube development
P.O. Box 26732, Tucson, spectra with special from Latham,
Arizona 85726, USA. handling of back- it is import-
ground ant to main-
tain operator
control, pro-
grams will be
changed to
interactive
Laboratorio di Astro- Gratton, L. processing of high IBM 370/125 16 k Fortran IV
fisica Spatiale, dispersion spectra, +
Frascati, Italy calculation of single Univac 1108
line intensities, p::
noise filtering for 2:
tTl
photo-electric
scanner spectra ~
Z
::t:
Center for Astrophysics Latham, D.W. reduces spectrograms CDC 6400 52 k s:
....
60, Garden Street N
and echellograms to tTl
Z
Cambridge, Mass. 02138 intensity versus
USA. wavelength, printer
3.3. a) One-dimensional time series (continued) ;g
o00
Institute Contact person Software Computer Core Remarks ~
is:
rn
plots, equivalent o'rj
widths. is:
(Program also used ~
and rewritten by =
~
Furenlid) 52o
Sterrewacht "Sonnenborgh" Rutten, R.J. Fourier restoration CDC 6500 48 k ALGOL
Zonnenburg 2, Utrecht, program ~
zo
the Netherlands t'!:I
~
Sterrewacht "Sonnenborgh" Rosenbaum, J. processing of stellar CDC 6500 -<
;.-
Zonnenburg 2, Utrecht, spectra for intensity z
the Netherlands versus wavelength, o
noise filtering 52
;J
~
Department of Astronomy Teske, R.G. determines equivalent interactive, ~
University of Michigan Stencel, R. widths, relative in use for ::l
Physics-Astronomy Bldg. Dopplershifts for two-dimension- ~
Ann Arbor, Michigan 48104 spectra lines from al scans of is:
USA. microdensitometer solar spectrum ~
;J
output plates ~
rn
Acknowledgements
REFERENCES
W.N. Brouw
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 301-307. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
302 W.N.BROUW
2. SOFTWARE REQUIREMENTS
3. MAP HANDLING
4. CLEAN
(\ ~-'" \J
\ ,-_...- /
o
declination
5. DISPLAY
REFERENCES
The main part of this paper has already been published {I}.
REFERENCES
de Jager/Nieuwenhullzen (eds.), Image Processing Techniques in Astronomy. 309. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
THE DESIGN OF VAMP SOFTWARE FOR THE MEASllREMENT AND REDUCTION OF
STELLAR SPECTROGRAMS
de Jager!Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 311-314. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
312 ROBERT J. RUTTEN AND GERARD P. VAN GELDER
ACKNOWLEDGEMENT
Drs. T. de Groot, Th. Gunsing and H. van de Stadt have taken part
in the definition of the new VAHP control program. Mr. J.H.G. Ro-
senbaum has written all the present reduction programs. We thank
Drs. T. de Groot, J. van Nieuwkoop, F. van de Stadt and C. Zwaan
for comments on the presentation of this paper.
DEVELOPMENT OF DIGITAL SYSTEM FOR STELLAR SPECTRA REDUCTION AT
TORUN OBSERVATORY
o
Polish Academy of Sciences, Institute of Astronomy,
Astrophysical Laboratory I
00
Institute of Astronomy of Nicolaus Copernicus
University, Torun, Poland
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 315-321. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
316 Z. TURJ:,O ET AL.
diversity of input data types and the requirements of the actual re-
duction problem on hand, software has been constructed in modular
form with a number of options enabling to select the range of the
reduction routines. To estimate the overall accuracy of our system
independently we compared our intensity tracings with those obtained
at the Dominion Astrophysical Observatory in Victoria for the same
plates. In another test, we compared equivalent widths derived with
our system with those published in the Solar Spectrum Atlas, Minnaert
et al. (1940). In all cases we have found quite satisfactory agree-
ments. Preliminary results obtained with this system, along with de-
tailed description of the hardware and testing have been published
elsewhere, Turlo (1974).
Examples of the practical application of the reduction system
described here are shown in Figures 1 and 2. Both spectra were con-
verted into intensity scale, convolved with the gaussian weighting
profile and normalized to the local continuum level.
SrlI
o
89 Her (F2Ia)
Tracings shown here were reproduced directly from the output tapes
obtained from the computer. The spectrum of 89 Her (F2 Ia) shown
in Figure 1 is in fact one of the first spectra taken with the
Canadian Copernicus spectrograph at the Torun Observatory. Figure 2
shows in detail the Ha line profile of the p Cas (F8 lap) spectrum
with the strong emission features clearly reproduced with our system.
DEVELOPMENT OF DIGITAL SYSTEM FOR STELLAR SPECTRA REDUCTION
317
t s;:l
r<.
;:l
P
0..-1
~
P
0
U
<1J
,.c:
~
0
~
"CI
<1J
N
0..-1
.....
cO
S
'0P"'
Po
cO
lS H
I
00
r...
Ul
cO
U
Q
4-<
0
<1J
.....
0..-1
4-<
0
a.
0 'Po"'
<Xl
LL
~
VI N
0
U
be
Or 0..-1
r...
o
318 Z. TUR:t,O ET AL.
B~ 51,"", ,pecl,um
Op. amplifiers
Analoque output
A'L ~ Ref. spectrum
Out
HD 25291 (FO nJ
nNlv
3 - rd difference
REFERENCES
I. D.W. Peat and A.C. Pemberton, The Observatory 2Q, 141, 1970.
2. L.B. Robinson, Proceedings of the Eleventh Colloquium of the
IAU Publ. Roy. Obs. Edinburgh §, 198, 1971.
3. W.D. Bonsack, Astronomy and Astrophysics 12, 374, 1971.
4. M.G.J. Minnaert, G.F.W. Mulders and J. Houtgast, Photometric
Atlas of the Solar Spectrum, Amsterdam, 1940.
5. Z. Turio, Bull. Astr. Obs. Torufi ~~, 9, 1974.
EVALUATION OF RASTER SCANS OF PHOTOGRAPHIC
STELLAR IMAGES x)
Erik H¢g
Astronomisk Observatorium, Brorfelde,
4340 T¢ll¢se, Denmark
INTRODUCTIOH
The fast devices becoming available for producing
raster scans of photographic stellar plates call for fust
numerical evaluation procedures to be used in an on-line
computer. The larger number of star images evaluated en-
creases the need for a procedure which is accurate not
de Jager/Nieuwenhuijzen (eds.). Image Processing Techniques in Astronomy. 323-328. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
324 ERIKH0'G
'"
r--
12
'-- 5 =2 ,.-
rt 12
6 5 =1 10 11 12
5=3 5 6 7 8 9 7\ 8
I....-
3 4 5 6 5 6 7 118
2 3 3 4
5=0 Vr =8
5=4 k=O 1
,
2 3 4
3, 4 =k
-
rL
, 22~5
,5 ,
5=5 r-- L......-
5=7 ),
5=6
- -
,
.
iJm (.)
25 ~ ~ ,
®, ....
d SO -d4O 20 I- " _<t').. ---o
o-_~•• "-00'0
0
15 I- -
a , I
/
units
1200 I- • side images -
o central
-
/y
~
1100
1000
: l~lm 0,
fi-
... (.)
//
I- 0
o~
900 I-
, -
:/
/
-
800 -
-
b star N° 8 4 10 7911 S 5
I I
-
.
dR-d U 5
0 (oi_ o _
0
C). 0" .- 00 0
-5 I- -
c , I I I I
x R
pm l- U
I>. -
o L
5 I- star S -
dE -d W 0 ~x - ~ t(t.)f
-
A
- ~
-5 -
-
(X)
-10 -d , I I (0) I
40 60 80 100 120 140 iJm dSO
de Jager/Nieuwenhuijzen (eds.). Image Processing Techniques in Astronomy. 329-334. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company. Dordrecht·Holland.
330 T. B. McCORD ET AL.
an image.
AVG Produces an image which is the average
of two or more images.
CONTR Draws an intensity contour map of an 5
image.
DELETE Deletes an image from the image file. <1 .01
DISP Scales (linearly or logarithmically) 2 .5
and displays an image in anyone of
six sectors on the terminal screen.
DIST Draws a graph of the distribution
of intensities in an image.
DIV Produces an image which is the quo- 15 2
tient of two images. Shifts the
numerator image vertically and hori-
zontally with respect to the denomi-
nator before dividing, so that iden-
tical features overlay each other.
DIVC Divides every element of an image 2 .25
by a constant.
ERASE Erases any area of the display 2 .01
screen.
FLAT Applies a flat field correction to 15 .25
an image.
FLUX Computes star intensity. 10 .5
FFT Computes the Fourier or inverse 30 15
Fourier transform of an image.
HELP Displays a description of the syntax, 2 .01
operands and function of any command.
HPFILT Applies a line-by-line high pass fil- 15 2
ter to an image. Used to emphasize
local features in an image at the
expense of overall shading variations.
LISTD Display's the directory to the images 5 .01
in the image file.
LOC Displays the screen coordinates of 1 . cn
point on the screen pointed to by
the user, using the terminal cursor.
LOGOFF Terminates the interactive system <1 .01
program.
MATCH Given the location of a feature in 15
one image, finds the location of the
same feature in a second image.
MULTC Multiplies each element of an image 2 .25
by a constant.
PEND Terminates process mode for currently 1 .01
active image. See PROC command.
PROC Starts process mode for an image. 2 .5
The image is copied into a temporary
disk file. This copy is called the
active image. All subsequent commands
IMAGE PROCESSING SYSTEMS DEVELOPED AT MIT 333
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 335-340. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
336 P. A. WAYMAN AND M. J. STIFT
2. PRECEDING WORK
The Magellanic Clouds are crowded regions where the effective back-
ground density is high and strongly variable across the plate. Hence
the difficulties in extracting accurate data on the brightness of
faint stars are severe. We have to be aware of the inherent "inversion"
presented by the task of measuring black stellar images, when what
we are aiming at is the brightness of a point of light, but perhaps
one so faint that it is barely distinguishable above the sky back-
ground. With an iris-diaphragm photometer, the brightness level can
be adjusted so that readings are taken with the diaphragm tightly
surrounding a stellar image. If this is done, the influence of the
HANDLING OF 'GALAXY' PHOTOMETRY 337
variations in sky background is diminished. Much of the information
in the bright images is wasted, but this wastage can be tolerated
when the aim is to maintain a certain accuracy expressed in magni-
tude (i.e •• light-ratio) and when it is possible to keep optimum
measuring conditions applicable to the difficult faint images.
The "tight" diaphragm helps to ensure a reasonable independence
with respect to varying sky-background, although Butler (1972a)
found a method of incorporating background density into his
least-squares reduction-process.
Methods of analysis and procedure that have proved useful for iris-
diaphragm measurements can have parallels whert considering the
choice of library profiles and weighting function with 'Galaxy'
machine (Walker, 1971). But complete adaptation to the requirements
of individual plates is not readily determined during measuring
runs. Thus there is a natural tendency to adopt one set of library
profiles covering, in 1024 steps, a large magnitude range, and
one may find in practice that almost all measurements of interest
come between, say. M = 100 and M = 300, and that one step in the
scale of ~ corresponds to 0.03 mag. This is just adequate, even
though a more expanded measure would sometimes be helpful. It was
found that there was a difficulty in setting the weighting-function
limits adequately far in because the same setting alters the amount
of light available to actuate the servo controls of the centering
mechanism. If the weighting-function is well into the central parts
of the stellar images, as is helpful in producing less dependence
on background fog (see Figure I), the diminished amount of light
available for the servos may result in a mechanical instability
during the centering process, causing the machine to "loop".
More sophisticated versions of 'Galaxy' should incorporate greater
flexibility in the setting of the weighting function when making
the photometric match with the library profiles.
T ot--__. .;.,. A
r
Fig, 1. an image with the library profile.
~,latching
(a) Library profile. (b) Stellar image profile
with normal background. (c) Image profile of
similar star with enhanced background, A is the
weighting function limit for maximum information,
B is preferable if we seek to avoid the influence
of background fog,
338 P. A. WAYMAN AND M. J. STIFT
4. REDUCTION PROCESS
where M = (l/p).(M/lOO)P
c
This form of M·1 was arrived at from the numerical integrations which
represented the matching procedure. The exponent p, which was found
in this way to lie in the range 2 to 3, represents the rather sudden
onset of background fog difficulties as M becomes smaller for the
faintest images.
After a few trials, it was found that this procedure was satisfac-
tory for the whole series of plates if p=2,S. The fitting of the
curve to the individual points used a rejection criterion
o = 0.25 mag. for excluding stars from the solution. and a series
of successive approximations guided by the computer operator
converged rapidly to a definite solution,
5. RESULTS OBTAINED
... .: ..
.- ... ...
,
-
. . ...::.,....
.......
12 ~
. .-... -.-.......
. • ..
\
~
-. .. . ......
.... .
.
1~ - • -d. • •
... s .. ......... .
,
. .. ., ,,
,.
- .
'"
-
$"
16 "'
I
REFERENCES
J. Dommanget
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 343-345. All Rights Reserved.
Copyright © 1975 by D. Reidel PUblishing Company, Dordrecht·Holland.
344 J. DOMMANGET
de Jager/Nieuwenhuijzen (eds.), 1I1UIge Processing Techniques in Astronomy. 347-352. All Rights Reserved,
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
348 D. A. KLiNGLESMITH AND E. DUNFORD
2. INSTRUMENTAL EFFECTS
x = A + B A + Cm + DmA
y =E+ FA + Gm + HmA
]
IMAGE 1-
ERRORS NOISE
comes from the spacecraft, at which time he and the night assist-
ant will be able to determine if all or part of the data should
be reduced or not. Obvious reasons for not reducing the data
might be over- or under-exposure, transmission error problems,
wrong star or any of the many other types of problems that we are
all familiar with at a ground based observatory.
REFERENCE
1. INTRODUCTION
< >
T, Tv T,Tv
from MIDI from
stellar spectra sensitometer
T 1
Reduce to Reduce to
Baker's density Baker's density
I
Record Compute the Plot the
INO) ~ normalized intensity H Atlas
tape IN (},)= I(,l.)/ IcO.)
~'Ol
~
.-:::;::; ~ 1'01
(~ ",-""
O'DIOt'
l-J ~ 6'60
'\1 ~- .""'-"-.
,-..
-.... ~I r-.......
8'60
.,.....,
--
L'eO
l.-/~ ~
9'60
~ lJ5 g'60
-~
~
'-"
1
~'60
........ ~
£'60
E-!
~. 0-. p,.
~ "0
-.; '::::-". U
~'60
~D
"t:l
C
j;:' CI!
1'60
c i--- ~'"-
I-- 4-l
0 I
=-
O·~
CI!
1-1
I
~"') S'60
.u
U
I
i6"'"
......, ~ ~ I
Q)
p,.
8'80 til I
Y~ I
Q)
..c:
L'BO .u I
~ b> 4-l
0
I
cJ.--- J::.?-7
9'80
I
c0
/" g'80 .,-1
I
'-'
::sCI!
( ~
.u
1-1
0 E-!
f'-. ),)
~'BO
p,.
<t! ;>-
£'80
~ ,
':::""-.
~'80
.
1'\ -.....;1\t.".
N
1'80 00
~ ~~ • ...1
~
?)
O·~
<::.:
~
-- ..,.. S'LO
~.
~
~ ?" ,'LO
'-= )/~ V
/" L 'LOO~
... .,
~. '"
o·
III
o· o o
CD
o· --.
o 'o" 0
0
356 A. CASSATELLA ET AL.
W f r(A) dA ( 1)
where
r 1 - I(A) (2)
- h.(A-A.)2
~ ~
rCA) l:. A. e
~ ~
- h.(A.-A.)2
J
r. Z. A. e
~ ~
(3)
J J ~
{"i. ~
~
Aik x.~ +L~ B'~ k y.~
Cik x. + L D' k y.
~ ~ ~ ~
Pk
Qk
(5 )
(6)
and
(7)
- h.(o)(A.-A.)2
X. r. - LA.(o)e ~ J ~ (8 )
J J ~ ~
d Aik
d h. (0)
~
d Aik
(9)
d h (0)
k
d 2 Aik
d h. (0) d hk(O)
~
h(O)h(O)
i k
h (0) h (0)
i + k
I:. e
J
hi (0) hk (0)
hi (O)+hk (0)
'" e ( 10)
hi (0) hk (0)
hi (0) + hk (0)
(II)
The expressions for the Bik' Cik and Dik are then easily
obtained from (9) and (II).
Even with a moderately large window the solution of the system
(5), is rather cumbersome; for instance for a 50 A wide window,
N = 2500, M '" 300 in our spectra. Fortunately there is an enormous
simplification due to the fact that Aik 7 0 very quickly when
IAk-Ail increases; in other words a l~ne is disturbed only by the
nearby lines.
Clearly, the dominant terms of the xi' Yi are obtained by
solving the "diagonal" systems
j A..~~
x. + B.. y.
~ ~~ ~
P.
~
(12)
t c ..
~~
x. + D.. y.
~ ~~ ~
R.
~
(0) (0)
If we call x.
~
, y.
~
the solutions of (12) or
360 A. CASSATELLA ET AL.
(0)
D.. P.
u. ~
- B.. Q.
~~ ~
x.
~ A.. D. - C.. B..
{
~~ ~ ~~ ~~
(13)
(0)
A.. Q. - C .. P.
y.
~
~~ ~
A.. D. - C .. B..
~~ ~
,
~~ ~ ~~ ~~
( 0)
x. x. + E;.
~ ~ ~
(14)
( 0)
y.
~
y.
~
+ n·
~
,
.
The initial values A. (0) and h. (0) are not critical and
. ~ ~
c an be eas~ly est~mated from the spectra; even better one may use
for them the values previously found from another star if avail-
able.
3. SATURATED LINES
w.
~
-~ A. ,
""""i ~
(15)
but for blended lines this is certainly not true and it ~s not
SEMI-AUTOMATIC DETERMINATION OF EQUIVALENT WIDTHS 361
very easy to find how the areas under the single profiles combine
together to give the combined profile. We tried therefore a
different kind of expansion, which worked rather satisfactorily and
is now currently used for processing our spectra.
For this procedure, we start from the well known empirical
formula for the profile of a line which was proposed by Minnaert
and by Unsold (1955) and is employed in the theory of the curve of
growth,
r (A) (16)
yeA) ( 17)
v
D A , ( 18)
c
- (A-A.)2 / D2
~
n Ae ( 19)
D = 0.13A
362 A. CASSATELLA ET AL.
and therefore equation (19) can be used until about 0.4 or 0.5 A
from the center of a line. This means that with very few exceptions,
like the 4227 Ca I line and another dozen of very strong lines,
all the lines in our spectra can pe represented by equations (16),
(17) and (19) even when the core is already strongly saturated.
Now, if at a certain wave-length A, we find that several
lines contribute to the absorption, it seems reasonable to sum all
the corresponding K£ -and write
11 (A) L A. e (20)
~ ~
r. (21)
J
L:. A. e
~ ~
r. ' (22)
J
r. R
J c
r. ' L A. e (23)
J 1 1
1,0
0,9
0,8
0,7
.
0,6
0,5
0,4
0,3
0,2
.
0,1
0,0
0 o
0 o
,
co..: co
co
co
aj
co.
N
0 0 o'It o
'It 'It 'It
(e) Another point which was carefully evaluated was that of filter-
ing the observed spectra for plate noise and deconvoluting for
the instrumental broadening. This can be done numerically
without complicating too much the processing program. But eventual-
ly we decided to use the raw spectra, without filtering and decon-
voluting them, on the following grounds.
The method of least squares in itself takes into account
accidental errors due to the noise; in fact for each line we
average upon a "theoretical" profile and fast oscillations due to
the noise are eliminated. But the use of theoretical profile
implies that the observed profiles be corrected for instrumental
broadening.
However, the use of a transfer function corresponding to some
instrumental profile would be equivalent to filtering in the
opposite way, that is to increasing the amplitude of the high-
fr~quency Fourier components, which are precisely those which are
more affected by noise. Also it is very doubtful whether it is
correct to consider the resulting profile as a simple convolution
of the "real" spectrum with the instrumental response, in the case
of a non-linear detector like the photographic plate.
On the other side (Gratton et al. 1975) the instrumental
broadening in our spectra was found to be of the order of + 0.045 A;
this is not negligeable, but is considerably less than the expected
Doppler half-width, 0.13 / J2 0.090 .
With a middle-size computer (an IBM 370/125) the processing of
a 35 A window containing about 200 lines takes about 15
minutes, including the printing of the result and the plot of the
observed and computed spectra (double precision is necessary). With
a larger computer this time is considerably shortened; also wider
windows can be used. In principle it is possible to process the
whole spectrum at once.
Our program in FORTRAN may be made, of course, available to
everybody who might be interested in them.
REFERENCES
1. INTRODUCTION
de Jager/Nieuwenhuijzen reds.), Image Processing Techniques in Astronomy. 367-372. All Rights Reserved.
Copyright @ 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
368 A. CASSATELLA AND R. VIOTTI
* This method was also used to. improve the photometric calibration
of some of the Hyades plates.
SPECTROPHOTOMETRY WITH OBJECTIVE PRISM PLATES 369
.2r---------------------------------------------------~
J
.1
-1
4 5 6 8
-2.5 LOG F1
4.78
5'"
y
3 2
FIG. 2. The oomputed energy speotrum of BD+15°637.
370 A. CASSATELLA AND R. VIOTII
+2r---~~------------------------------------------~
UBGYRI
+ J;«. ... 0 • 0
+1
-1 Y1016 CYG
+
-2 m
11 13
line blocking
6
.4 -
*
y Tau
,4-
.2 - Y) Cep
28
96
178
... HYADES
~
REFERENCES
Ardeberg~ A. and Virdefors, B.: 1972, Astron. and Astrophys. ~,
177 •
Baratta, G.B., Cassatella, A. and Viotti, R.: 1974, Astrophys.J.
l§1,65 1 •
Cassafe1la, A., Maffei, P. and Viotti, R.: 1973, in Speotral Clas-
sifioation and Multiootour Photome~ry, ,ed. by Ch. Fehrenbaoh
and B.E. Westerlund, TI. Reidel Pub. Co., TIordreoht, p. 127.
Cassatella, A•. and Viotti, R.: 1972, Rapporto di Rioeroa, Osserva-
torio Astronomioo di Roma, N.3.
Cassatella, A. and Viotti, R~: 1974, R~p~orto Interno, Laboratorio
Astrofisica Spaz~ale, Frascati, N.5.
Nandy, K.: 1964, Publ. Roy. Obs. Edinburgh,..1, 142.
Nandy, K. and Smriglio, F.: 1970, Observatory, 2Q, 114.
Oke, J.B. and Conti, P.S.: 1966, Astrophys.J. 143, 135.
Woszozyk, A., Smolinski, J., Maron, N., Strobe~A. and Krempeo,
J.: 1968, Astrophys ~ and Spaoe Soienoe.l, 264.
LE CENTRE DE DEPOUILLEMENT DE CLICHES ASTRONOMIQUES
LE MATERIEL
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 373-378. All Rights Reserved.
Copyright © 1975 by D. ReidelPublishing Company, Dordrecht-Holland.
374 M. AUVERGNE ET AL.
systeme evolue.
- de l'existence d'un logiciel permettant de traiter les
cliches les plus divers.
- de la possibilite offerte par l'interface de piloter la
platine simultanement en x et y.
Pour cela nous avons adjoint une unite de disque de 2.5 millions
d'octets, une unite de bande magnetique 7 pistes et une console
graphique Tektronix 40IO. Celle-ci joue un r8le fondamental dans
Ie traitement gr~ce a la rapidite des entrees sorties et a la
possibilite de tracer courbes, isophotes, ••• Nous l'avons equipee
d'un reprographe electronique afin de conserver, Ie cas echeant,
un document apres utilisation.
LE LOGICIEL
de cliches. Tous ces programmes ont ete ecrits sous forme conver-
sationnelle. telle que l'utilisateur dispose de nombreuses
possibilites. Ces premiers programmes permettent :
- de tracer sur l'ecran de la console une ligne ou portion
de ligne d'enregistrement, de determiner l'abscisse et la densite
de points choisis par un reticule electronique (CESAR~.
- de tracer des isophotes selon un code ternaire cyclique a
partir d'une zone de 1 'enregistrement, et de determiner les
coordonnees de points reperes (ISIS).
- de transformer les densites de tout un champ en intensite
I, logI, ou magnitude m au dessus du fond de ciel (CALIB~.
- de filtrer les mesures avec un operateur choisi par
l'astronome (VENUS).
- d'effectuer des operations arithmetiques entre des fi-
chiers (ARIFIL).
DIRECTIONS DE TRAVAIL
LE CE1'TRE DE DEPOUILLEMENT
CONCLUSION
1. Introduction
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 379-384. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
380 ROBERT D. CHAPMAN ET AL.
H-a.6563A
PHOTOGRAPH
\ ..
The x-ray and EUV observations were made with the GSFC
spectroheliograph flown on-board OSO-7. Details of this instru-
ment have been published by Neupert et al. (1974) and by Neupert
and Underwood (1974). The x-ray and~UV-observations consist of
raster scans of a 5 x 5 arc minute area of the solar disk made
using a 10 x 20 arc second aperture. Several OSO-7 rasters made
over a 15 minute time period were averaged together to produce
one raster in each of the observed wavelengths.
INTERCOMPARISON OF GROUND BASED AND SATELLITE PICTURES OF THE SUN 381
2. Discussion
CHROMOSPHEIE
References
1.- INTRODUCTION
This work has been performed to test the new and interesting re-
sults obtained previously with a photographic isodensitometric m~
thod about the photometric evolution of solar flares (Falciani et
al., 1972; Falciani and Rigutti, 1972 a,b) and to study the degree
of utility and reliability, and the general performances of high
speed, computer controlled devices in the photometric analysis of
extended sources. We studied some series of good Ha solar filter-
grams, obtained during 1969 (May 15-16-17-25 and Oct. 25-27), at
the Athens National Observatory (we warmly thank Dr. C.J. Macris
for having kindly put at our disposal such a material), with ti-
me resolution of about 30 sec, with uniform exposure and high ph~
tometric accuracy (F::I 5000 fittergrams).
To study the huge amount of the above material we used the S.A.D.
A.F. computer controlled flying-spot digitized machine of the Isti
tuto di Elaborazione dell'Informazione (Pisa). This instrument
has been extensively described elsewhere (Azzarelli and Panicucci,
1972; Carlesi and Montanari, 1973; Carlesi, 1975). It is a conven-
tional type of flying-spot photometer (see the general plan in
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 385-390. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
w
00
0\
IDP8/1
12 ~III IIId
8~ mI.",
1.51'1" Clcllll •• IIIEIFIC
110
Omol
"
~
I IXIS 1101 100Uum 10·1b~II.2
JDIITICK II COolOIUHS
r"
>
N
~
:;.;
ttj
,,
'"
'"
'"
ux.
lin xx
"Xl!
.....
xux
lI')(lCX'U",
lI .... lI'X'WlIlI),UWl
xn lIxrxnV.'lXII)(YXYXY)(X)(lXllll
'IIX . . . . . . . X)(lIX ..... 10'X ..... II' ... XXX'f.OIXX1XJX'lXll')ClIXlt'lI'XXJX'I'1'ltXlI)(IX)/lr.1)/)('1
leXX 'II X VlIXX It' X 11 X 1I' xY I( 'If)( llX It' xx X 'to')( 11' no x 'I X yx rx 'II kit' U X YX)( X It' Xli n x xx 'tX1I' )(Y)(lI')()c)C x x x x x X 1I' X xx XX)(o: X X)( X II X)( xxv x x x x x 'II 0')(
lIXl "OlIUX'IflIIOIXx1)(YlClIX'llXllXllllWXUlIXXXXXX)()(XXlI'XXXXl!XYXX)lIt'XXXYXVXY)()t')()(lI'lCXxxxxxxnOxxxuX'l/x.II'YX'tXy)('tXyx . . . . . y)('iJ'I'
"II XlI' '(xx )(XlI 'ill' It Xn: YXYX Y XlI)( lI'XY lI' lIXl/lllt)( X X.. X1l X)t' X X)( X)( Yx11 XlI' X11)( 'If lOt' X)( X'11')( r)('1')( XX)(1t' X)( X)( Y xn lfX)( x II X Xxv XlI')( on!)()( XIt' X'tX lI' xxxx Xy XY xli x'" XX
X Xl KXX Xli)( 11 XX X'II' Xnxll"l' IClI lUX X11)( n't x lI' X)( Xli' X'11')( xx lI' Xy XX)( XlI' X X11')( 10: YXli' X'II XlI X)()( 'Ox XX XY Xxx '0)' x n)( x yx't)( 'Ii X)! XlI' X'11')( 'tX lI' XIII XXy XIt' X_ lilt' xx. Y J
lOCI X)(lI'xnUJXlI'XYlClInx'tXlIXlIxnnll)'YXlt'XlI'X'tXlt'XltXltlCY)('()(lI'J(lt'XlI)(YXxXYXyxunXXXl(yxu't)(lI'XIt'XY)ClIYlI'lIY)(lIKXl.X
II xx ".IX lin Y)I" X.............. x .. X .... )(X ...... X .. J)' X'I')()I' x .... xu .... )(X ••• 'I' ....... rx.x 'I' 'I'll':
'I'll'" .... )()(n .... wx •• w.. 'I' .. 'I'x)(X ... ~.XlI' .. yXll'x ........ 'I' •• )(lI'x •• !()(lI'Xlt' ..
"lilt ... xx ... X)CUXIt'l'Xl!n .. yu.lt .. y.yXll' .. y ......... ltx
.. IX YX'I'xny .... x .. xll'ltWX'l'Xy .... ly .... ny'l' ..
XIX Xl'l'Xl!.""UlIuU •
.. n Ynx.nlllt
rn n.r ..
'"
The step filter for the densitometric comparison and test of the
machine has been calibrated in the same density scale D of the
calibration curve of the analyzed photographic emulsion.
Checkes for the photometric stability of the flying-spot gave re-
sults of '" 2'70. for D < 0.6, '" 8'700 for 0.6 ,.;; D < 1.3 and", 1 -;- 2'70
for D ~ 1.3 (the latter is due only to the difficulties to disen-
tangle and read exactly neighbouring hystograms at high D values).
We calibrated with extreme care the mean value of the scanning
spot area and tested the reproducibility of all the photometric
analysis (hard + soft) by scanning the same photogram many times
and with various working conditions during several days. We obta~
ned a rms error in the area determinations ranging from 2.3% to
7% for areas of 20 mm 2 and 0.3 mm 2 respectively.
To check the differences between this analysis method
and the photographic isodensitometric one, we measured many filte£
grams of our previous work (Falciani and Rigutti, 1972 a) and ob-
tained a rms agreement of the order of 2 + 10% and a small syste-
matic deviation of 1 -;- 2% in the mean values for 1.1 < j < 1.6.
Finally we like to stress how powerful and flexible in comparison
of similar ones is the present method of analysis, wbich supply
all the wanted data (areas, isophotes, points determinations, etc.),
keeps in a numerical matrix the information of the original phot~
grams, viz. in a form particularly suitable to any further elabo-
rations, and presents very simply any feed-back procedure between
the first approximation results and the next steps of the work.
REFERENCES
Gregory C. Alvord
State University of New York at Albany, U.S.A.
Daniel A. Klinglesmith, Lawrence Dunkelman
Goddard Space Flight Center, U.S.A.
Robert r~. Mercer
Dudley Observatory, U.S.A.
Abstract
The comparison of the vignetting function for Apollo 15, 16
and 17 35mm Nikon cameras is presented. A technique for remov-
ing this effect is described, and demonstrated on an image of the
L4 point of the Earth Moon system.
de Jager/Nieuwenhuijzen reds.), Image Processing Techniques in Astronomy. 391-396. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
392 GREGORY C. ALVORD ET AL.
obtain
obtain data vignetting image
image on on calibrated
cal ibrated film film
scan to sample
film image and
scale 0,255
in density
determine and
remove film
density fog
scan calibration
wedges to
obtain film
characteristics
rescale to Log
intensity of the
image
output image
subtract is vignetting
vignetting image 1-----1 removed
from data image Log I picture
(a) (b)
(c) (d)
(a) (b)
(c) (d)
4. examples of results
1. INTRODUCTION
de Jager/Nieuwenhuijzen (edll.), Image Processing Techniques in Astronomy. 397-402. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht-Holland.
398 R. BUCCHERI ET AL.
2. METHODS OF SELECTION
3. 1. Parameters used
In the following we give some prel iminary resul ts on the appl ica-
tion of a non parametric method of automatic selection to the
data obtained during a calibration run of the COS-B spark
chamber For information about COS-B spark chamber see
Bignami et al. (1974). Two classes of parameters have been used
in this application. The first already aoplied to the data obtained
from an ootical spark chamber by Buccheri et al. (1975), takes
into account combination of sparks in a section of the event
picture. This kind of parameter can be considered " ro bust ll
because its definition is independent on the particular structure
of the image and we can expect rather equivalent selection
efficiency when applied to different classes of data. The second
class of parameters used are "ad hoc ll , defined in function of the
doubl e track structure of the gamma ray event pi cture : for these
we can expect good selection efficiency only in very restricted
cases.
3. 2. The method
••
• • • • •
• ..
. ••
• •
·. • •
. • ·• • •
• · • • •
.• • .
•
• •
(f)
.....
c
~ OI--~44~-+nr-u+--wbr~r+~~~-.~--~~-nr+-'dr~~~~~
>.
L
co
L
.0
L
co
where N is the average value of the N.'s.ln this case the train-
ing hystogram has in abscissa intervals of values for £2 and
in ordinate the difference (Co. - Cj) between the frequency
of the If -events for which i:.. contained in a fixed interval i!
and the same frequency for the -events. y
- In the third phase the selection is performed by using a linear
decision function and a fixed threshold value ¢ according to
the following criterium { ~ ¢ ~ ,-
(dK - f
d < ¢ ~ F
where (dl'- d;J is the recorded difference between the frequency
of occurrence of the measured value of the chosen parameter
in the unknown event for I-events and the same value for ?f
4. DATA ANALYZED AND RESULTS OBTAINED
The sample of data used for the application of the method illust-
rated above is made of 891 events ( 434 If + 457 obtained t)
from an accelerator calibration run of the COS-B wire spark
chamber and kindly suppl ied by the "Caravane COllaboration".
Two examoles of event images are given in figs. 1 and 2. The
preprocessing ohase (not yet optimized) applied to the entire
sample has given as result the exclusion of 269 events of which
55?f and 214 ¥
The training hystograms have been derived for
0
I
/
Fig.4 I
I
Fig.5 fi
.f I I
, I
I
I
/
/
I
I
// g
'I /
®I
.~
, t'
,JP/
~
___ -.2 ." . ., (,
/
T ! !
Fig.6 Fig.7
"
I
.K
., I
I
I
.1, ,/
,/
fit ,/
,/
/(!!f /
,/
,/
Jlf
,.,.e/
•9 • .2. • '-t .6 4.. •
REFERENCES
INTENSITY
SPECKLED
BACKGROUND
PLANET
·O~Osec Arcseconds
Fig. 1 : Typical spread function to be expected with the
Large Space Telescope. A circumstellar planet is about
10 4 times fainter than the scattered light level.
Lock-in detection.
A B c
APPENDIX 1
APPENDIX 2
References
1. INTRODUCTION
de Jager/Nieuwenhuijzen (eds.), Image Processing Techniques in Astronomy. 411-418. All Rights Reserved.
Copyright © 1975 by D. Reidel Publishing Company, Dordrecht·Holiand.
412 w. F. VAN ALTENA AND L. H. AUER
In this paper we present results on the digital dissection
of images measured on the PDS microdensitometer. We address two
basic questions: 1) Is the PDS microdensitometer capable of doing
precision differential astrometric measurements and 2) Can the
numerical processing be done in reasonable amounts of computing
time?
2. MEASUREHENTS
3. IMAGE ANALYSIS
x = Lx.
y y
w(y) • N(y) / Ly w(y) • N(y) (2)
Px (x) 'V a
x
+ bx x + [Nx /(2TI) 1/2 a].
x c
2
x
2
exp[-0.5 (x - x ) /a ] (4)
The form of Eq. 4 follows directly from the model that D(x,y) is
the sum of a linear background and gaussian image; it holds even
if the image axes do not lie along the x and y coordinates.
-
o
C\J
-
o
co
~----------r---------~-----------r--------~
0- 80 . 00 -40.00 0.00 40.00 80.00
X
Fig. 1. Fit to an observed marginal distribution.
6. CONCLUSIONS
7. ACKNOWLEDGEMENTS
REFERENCES
Edited by
I. C. de Jager (ed.), The Solar Spectrum. Proceedings of the Symposium held at the University o{ Utrecht,
26-31 August, 1963. 1965, XIV +417 pp.
2. J. Ortner and H. Maseland (eds.), Introduction to Solar Terrestrial Relations, Proceedings o{the Summer
School in Space Physics held in Alpbach, Austria, July 15-August 10, 1963 and Organized by the Euro-
pean Preparatory Commission for Space Research. 1965, IX + 506 pp.
3. C. C. Chang and S. S. Huang (eds.), Proceedings of the Plasma Space Science Symposium, held at the
Catholic University of America, Washington, D.C., June ll-14, 1963. 1965, IX + 377 pp.
4. Zdenek Kopal, An Introduction to the Study of the Moon. 1966, XII +464 pp.
5. B. M. McCormac (ed.), Radiation Trapped in the Earth's Magnetic Field. Proceedings of the Advanced
Study Institute, held at the Chr. Michelsen Institute, Bergen, Norway, August 16-September 3, 1965.
1966, XII+901 pp.
6. A. B. Underhill, The Early Type Stars. 1966, XII + 282 pp.
7. Jean Kovalevsky, Introduction to Celestial Mechanics. 1967, VllJ +427 pp.
8. Zdenek Kopal and Constantine L. Goudas (eds.), Measure of the Moon. Proceedings of the 2nd Inter-
national Con{erence on Selenodesy and Lunar Topography, held in the University of Manchester, Eng-
land, May 30-June 4,1966.1967, XVIII +479 pp.
9. J. G. Emming (ed.), Electromagnetic Radiation in Space. Proceedings of the 3rd ESRO Summer School
in Space Physics, held in Alpbach, Austria,from 19 July to 13 August, 1965.1968, VIII + 307 pp.
10. R. L. Carovillano, John, F. McClay, and Henry R. Radoski (eds.), Physics of the Magnetosphere,
Based upon the Proceedings olthe Conference held at Boston College, June 19-28, 1967. 1968, X + 686 pp.
II. Syun-Ichi Akasofu, Polar and Magnetospheric Substorms. 1968, XVIII + 280 pp.
12. Peter M. Millman (ed.), Meteorite Research. Proceedings ofa Symposium on Meteorite Research, held
in Vienna, Austria, 7-13 August, 1968. 1969, XV + 941 pp.
13. Margherita Hack (ed.), Mass Lossfrom Stars. Proceedings of the 2nd Trieste Colloquium on Astro-
physics, 12-17 September, 1968. 1969, XII + 345 pp.
14. N. D'Angelo (ed.). Low-Frequency Waves and Irregularities in the 10nosphere. Proceedings of the 2nd
ESRIN-ESLAB Symposium, held in Frascati, Italy, 23-27 September, 1968. 1969, VII + 218 pp.
15. G. A. Partel (ed.), Space Engineering. Proceedings ol the 2nd International Conlerence on Space Engi-
neering, held at the Fondazione Giorgio Cini, Isola di San Giorgio, Venice, Italy, May 7-](}, 1969. 1970,
Xl +728 pp.
16. S. Fred Singer (ed.), Manned Laboratories in Space. Second International Orbital Laboratory Sympo-
sium. 1969, XIII + 133 pp.
17. B. M. McCormac (ed.), Particles and Fields in the Magnetosphere. Symposium Organized by the Summer
Advanced Study Institute, held at the University of California, Santa Barbara, Cali{., August 4-15,1969.
1970, Xl + 450 pp.
18. Jean-Claude Peeker, Experimental Astronomy. 1970, X + 105 pp.
19. V. Manno and D. E. Page (cds.), Intercorrelated Satellite Observations related to Solar Events. Pro-
ceedings of the 3rd ESLAB/ESRIN Symposium held in Noordwijk, The Netherlands, September 16-19,
1969.1970, XVI+627 pp.
20. L. Mansinha, D. E. Smylie, and A. E. Beck, Earthquake Displacement Fields and the Rotation of the
Earth. A NATO Advanced Study Institute Conference Organized by the Department of Geophysics,
University of Western Ontario, London, Canada, June 22-28, 1969. 1970, XI + 308 pp.
21. Jean-Claude Peeker, Space Observatories. 1970, XI + 120 pp.
22. L. N. Mavridis (ed.), Structure and Evolution of the Galaxy. Proceedings (){ the NATO Advanced Study
Institute, held in Athens, September 8-19,1969.1971, VII+312 pp.
23. A. Muller (ed.), The Magellanic Clouds. A European Southern Observatory Presentation: Principal
Prospects, Current Observational and Theoretical Approaches, and Prospects for Future Research. Based
on the Symposium on the Magellanic Clouds, held in Santiago de Chile, March 1969. on the Occasion
of the Dedication of the European Southern Observatory. 1971. XII + 189 pp.
24. B. M. McCormac (ed.), The Radiating Atmosphere. Proceedings of a Symposium Organized by the
Summer Advanced Study Institute, held at Queens University, Kingston. Ontario, August 3-14, 1970.
1971, XI+455 pp.
25. G. Fiocco (ed.), Mesospheric Models and Related Experiments. Proceedings ql'the 4th ESRIN-ESLAB
Symposium, held at Frascat;, Italy, July 6-/0,1970. 1971, VIII +298 pp.
26. I. Atanasijevic, Selected Exercises in Galactic Astronomy. 1971, XII + 144 pp.
27. C. J. Macris (ed.). Physics of the Solar Corona. Proceedings 01' the NATO Advanced Study Institute
on Physics of the Solar Corona, held at Cavouri- Vouliagmeni. Athens, Greece, 6-17 September 1970.
1971, XII+345 pp.
28. F. Delobeau, The Environment of the Earth. 1971, IX+ 1 \3 pp.
29. E. R. Dyer (general ed.), Solar-Terrestrial Physics/1970. Proceedings qj'the 1nternational Symposium
on Solar-Terrestrial Physics, held in Leningrad. U.S.S.R., 12-19 May 1970.1972, VIII +938 pp.
30. V. Manno and J. Ring (eds.), Infrared Detection Techniques for Space Research, Proceedings of the
5th ESLAB-ESRIN Symposium, held in Noordwijk, The Netherlands, June 8-11, 1971. 1972, XII + 344
pp.
31. M. Lecar (ed.), Gravitational N-Body Problem. Proceedings ql'IAU Colloquium No. 10, held in Cam-
bridge, England, August 12-15, 1970. 1972, XI+441 pp.
32. B. M. McCormac (ed.), Earth's Magnetospheric Processes. Proceedings of a Symposium Organized by
the Summer Advanced Study Institute and Ninth ESRO Summer School, held in Cortina, Italy. August
3D-September /0. 1971. 1972, VIII +417 pp.
33. Antonin Riikl, Maps q( Lunar Hemispheres. 1972, V +24 pp.
34. V. Kourganoff, Introduction to the Physics of Stellar Interiors. 1973, XI + 115 pp.
35. B. M. McCormac (ed.), Physics and Chemistry of Upper Atmospheres. Proceedings q( a Symposium
Organized by the Summer Advanced Study Institute, held at the University q( Orleans, France, July
31-August 11. 1972.1973, VIII + 389 pp.
36. J. D. Fernie (ed.), Variable Stars in Globular Clusters and in Related Systems. Proceedings of the IAU
Colloquium No. 21. held al the Unit'ersity of Toronto, Toronto, Canada, August 29-31, 1972. 1973,
IX+234 pp.
37. R. J. L. Grard (ed.), Photon and Particle 1nteraction with Surfaces in Space. Proceedings ql'the 6th
ESLAB Symposium, held at Noordwijk, The Netherlands, 26-29 September, 1972. 1973, XV + 577 pp.
38. Werner Israel (ed.), Relativity, Astrophysics and Cosmology. Proceedings of the Summer School. held
14-26 August, 1972, at the BANFF Centre, BANFF, Alberta, Canada. 1973, IX+323 pp.
39. B. D. Tapley and V. Szebehely (eds.), Recent Advances in Dynamical Astronomy. Proceedings of the
NATO Advanced Study Institute in Dynamical Astronomy, held in Cortina d'Ampezzo, Italy, August
9-12.1972.1973, XIII +468 pp.
40. A. G. W. Cameron (ed.), Cosmochemistry. Proceedings of the Symposium on Cosmochemistry, held at
the Smithsonian Astrophysical Observatory, Cambridge, Mass., August 14-16, 1972. 1973, X + 173 pp.
41. M. Golay, Introduction to Astronomical Photometry. 1974, IX + 364 pp.
42. D. E. Page (ed.), Correlated Interplanetary and Magnetospheric Observations. Proceedings qj'the 7th
ESLAB Symposium, held at Saulgau, W. Germany, 22-25 May, 1973. 1974, XIV +662 pp.
43. Riccardo Giacconi and Herbert Gursky (eds.), X-Ray Astronomy. 1974, X+450 pp.
44. B. M. McCormac (ed.), Magnetospheric Physics. Proceedings of the Advanced Summer Institute, held
in Sheffield, U.K., August 1973.1974, VII +399 pp.
45. C. B. Cosmovici (ed.), Supernovae and Supernova Remnants. Proceedings of the International Confer-
ence on Supernovae, held in Lecce, Italy, May 7-Il, 1973. 1974, XVII + 387 pp.
46. A. P. Mitra, Ionospheric Effects of Solar Flares. 1974, XI + 294 pp.
49. Z. Svestka and P. Simon (eds.), Catalog of Solar Particle Events 1955-1969. Prepared under the
Auspices q( Working Group 2 of the Inter-Union Commission on Solar-Terrestrial Physics. 1975,
IX+428 pp.
50. Zdenek Kopal and Robert W. Carder, Mapping of the Moon. 1974, VIII +237 pp.
51. B. M. McCormac (ed.), Atmospheres of Earth and the Planets. Proceedings of the Summer Advanced
Study Institute, held at the Unil2!rsity q( Liege, Belgium, July 29-August 9, 1974. 1975, VII +454 pp.
52. V. Formisano (ed.), The Magnetospheres of the Earth and Jupiter. Proceedings of the Neil Brice
Memorial Symposium, held in Frascat;, May 28-June 1,1974. 1975, XI+485 pp.