You are on page 1of 1

Accuracy and Quality Comparison of Orthorectified High Resolution Satellite Images

Panagiotis AGRAFIOTIS , Andreas GEORGOPOULOS


Laboratory of Photogrammetry, School of Rural and Surveying Engineering, National Technical University of Athens pagraf@central.ntua.gr drag@central.ntua.gr

SCOPE
Accuracy and quality assessment of orthorectified high resolution satellite imagery from the Pleiades 1B and the Geoeye-1 data through : Radiometric quality assessment Comparative evaluation of their quantitative and qualitative properties Investigation of the advantages and limits of the Pleiades B1 Imagery for producing Large Scale Orthophotos (LSO)

CONCLUSION
Metric and radiometric comparisons proved that the Pleiades B1 high resolution satellite imagery is capable of producing highly accurate and reliable Large Scale Orthophotos (LSO) for mapping and GIS applications having clear advantage against similar products.

STUDY AREA
Greece Cyclades The small island of Antiparos
Area 35000 km2

QUALITY AND ACCURACY ASSESSMENT


IMAGE QUALITY ASSESSMENT

Pleiades B1 stereopair (0.50m GSD) from April 2013 Two adjacent GeoEye-1 stereopairs (0.50m GSD) from December 2011 Aerial orthophoto (LSO) provided from Ktimatologio S.A. (Hellenic Cadastre) (0.50m GSD) from 2007

TEST DATASET

Pleiades B1

Geoeye-1

LSO

Visual inspection of orthophotos for errors and/or image defects Geoeye-1 image : Abrupt change of brightness and contrast Many regions presented extremely dark/light tones, predominantly on the buildings and along the roads Radiometric saturation of certain regions is especially prominent Pleiades B1 image : Presents much more information and clearer forms Is much darker than LSO and Geoeye-1 imagery

CONTROL DATASET
Aerial orthophoto (VLSO) provided from Hellenic National Cadastre & Mapping Agency S.A. (NCMA S.A) (0.25m GSD) from August 2012
16 GCPs distributed evenly on the island area GPS observations Fast static method GCPs mean uncertainty of 0.010 m
Panchromatic GSD 0.50m Multispectral GSD 2.00m Pansharpened GSD 0.50m

METHODOLOGY

GPS measurments Pansharpening HSV for Pleiades data and HCS for Geoeye data
Photogrammetric georeference (LPS and Photomod 5)

SPATIAL ACCURACY

+
Pleiades Distribution GCP Check Point

=
Pleiades B1 4 6 0.045 Geoeye 1 5 10 0.160 1st degree polynomial RPC refinement

where, d is the deviation n the number of check points

X,Ycheck the check points coordinates measured on control dataset and X,Ydata the points coordinates measured on test dataset

Data GCPs Check Pts RMSE (m)

RESIDUALS Data source Check Pts StDEV (m) RMSEX,Y (m) RMSEX,Y (m) Pleiades B1 25 0.374 0.389 0.522 Geoeye-1 LSO Pleiades B1 26 26 25 0.388 0.467 0.658 0.607 0.495 0.838 AFTER BIAS REMOVAL 0.381 0.578 0.6455 Geoeye-1 26 0.415 0.453 0.453 LSO 26 0.423 0.465 0.465

DEM extraction Orthophoto production Independent measurements for checking horizontal accuracy of orthoimages

Automatic procedure for Pleiades B1, semi-automatic for Geoeye-1 GCPs used for transferring the scale and real elevation values 10 m elevation interval Bilinear Convolution GSD 0.50 m Orthorectification of all stereopairs images for DEM checking Check Points measured on VLSO Distributed according to JRC Guidelines for Best Practice and Quality Checking of Ortho Imagery and NSSDA standards Areas with different terrain relief and land cover included 25 Check Points for Pleiades B1 26 Check Points for Geoeye-1 and LSO Check Points Distribution

HR RESIDUALS Data source Pleiades B1 Geoeye-1 LSO

The BIAS is estimated by the formula:

RMSEXY (m)

0.646

0.866

0.649

Where, is the random error

In LSO and Geoeye-1 RMSEY StDEVY indicating a systematic error in Y axis Geoeye-1 data have the bigger RMSEXY possibly due to bad image quality leading to erroneous GCPs measurements Pleiades B1 orthoimage has almost the same external accuracy as the orthoimage from aerial imagery (LSO from Ktimatologio S.A.) All data are adequate for producing LSO, according to JRC and NSSDA accuracy standards