The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XXXIX-B4
https://doi.org/10.5194/isprsarchives-XXXIX-B4-139-2012
https://doi.org/10.5194/isprsarchives-XXXIX-B4-139-2012
31 Jul 2012
 | 31 Jul 2012

SATELLITE 'CANOPUS-V' IMAGE PROCESSING TECHNOLOGY DEVELOPMENT FOR CARTOGRAPHY PURPOSES BASED ON PRELAUNCH SIMULATION

V. V. Nekrasov and E. Makusheva

Keywords: Cartography, Processing, Adjustment, Orthorectification, Sensor, Georeferencing, Satellite, Accuracy

Abstract. “Canopus-V” satellite payload consists of two cameras – one of them panchromatic with 2.1 m GSD, another multispectral with 10 m GSD. Navigation system accuracy provides georeference accuracy about 70-80 m without using of control points. The peculiarities of high-resolution panchromatic camera built with micro-frames technology requires to design surveying technology processing for obtaining continuous coverage. Automatic technology of searching tie points in overlapped areas of micro-frames was developed by authors. Block adjustment is done automatic. Rational polynomial coefficients (RPC) calculates after block adjustment for using in standard photogrammetric software. For “Canopus-V” sensors was developed mathematic model and software package for RPC calculating. Searching for tie points use combination of feature and area matching algorithms in overlapped areas of micro-frames. Process of satellite sensors shooting was simulated with use of calibration data (orthophoto, digital elevation model – DEM, control points for test area used as check points). Subpixel accuracy was achieved for block adjustment. Accuracy for resulting orthophoto was checked with control points and was 35-55 m. Using of two control points for block increases resulting orthophoto accuracy two times. Simulation results and accuracy estimation shows that “Canopus-V” panchromatic camera data can be recommended to modify maps for the scales 1:25000. “Canopus-V” multispectral camera data can be used to modify maps for the scales 1:100000.