You are on page 1of 8

Globale Operatoren: 2D Fouriertransformation

usw.
Looks the same!
The lens makes a
“FFT”

f… focal length of Screen at the


Parallel
the lens pupil plane
“rays”
(=back focal
position of the
first lens)
This is no intermediate
4F-System Manipulation of numerical image! It looks like,
Aperture or Pupil Function, here because we are imaging
a periodic object

Image

f… focal length of the


lens

2d Fourier Transform 2d inv. Fourier Transf.


Fourier Transform is a collection of scalar products = dot products

If norm of a is 1, i.e. ||a|| = 1, then the


dot product is the signed length of
the orthogonal projection of b on a

a with ||a||=1

Dot product says: how much of b is in a

Extend this to functions u(x) and v(x): How much of v(x) is in u(x) ?

Hence, the sum in a • b turns into an integral for u • v:

For the complex


valued case:
Fourier transform is just a scalar/dot
[F f](ω)= product of a function with parametrized
basis oscillations
Parametrized basis oscillations? In our case: n=2 and x and w (= ω )
are 2-dim vectors: w=(w_1,w_2) and
x=(x_1,x_2)
= exp(-iwx)=exp(-i[w_1x_1+w_2x_2])

=cos(w_1x_1+w_2x_2) - i sin(w_1x_1+w_2x_2)

x = (x_1,x_2) are spatial coordinates, vanishing in the integral for [Ff](w)


w = (w_1,w_2) are the parameters of the oscillation

In a diagram with axis w_1 and w_2, basis oscillations are just a point!
Mostly, this point and the point for the conjugated oscillation are shown.
The resulting image shows the absolute values of the Fourier transform.

Now review the slides 2 and 3

An imaging optical system “performs” a 2D FT of the object to the energy


distribution in the pupil plane, where the aperture stop is located.
Then it inversely 2D FT’s this distribution in the pupil plane to form an image
energy distribution. A sensor would collect intensities of the energy: the square of
the absolute value of the complex wave front, i.e. the square of the amplitudes.
All aberrations and aperture stop effects can be described as manipulations of
the pupil plane – i.e. the 2D pupil function. See slide 5
To understand how an aperture stop works, we
need it’s inverse FT. This is just the normal FT
on the [Ff](w), but with , i.e. the complex
conjugated basis functions.

What if the aperture stop is an infinite slit in y


and with limit interval [-a,a] in x? Let’s compute
it: Aperture stop (e.g. slit) here
Optical system, front view

a
Slit stop – no
limit in y, [-a,a]
f f limit in x

You might also like