Effects of image compression and illumination on digital terrain models for the stereo camera of the BepiColombo mission

The great amount of data that will be produced during the imaging of Mercury by the stereo camera (STC) of the BepiColombo mission needs a compromise with the restrictions imposed by the band downlink that could drastically reduce the duration and frequency of the observations. The implementation of an on-board real time data compression strategy preserving as much information as possible is therefore mandatory. The degradation that image compression might cause to the DTM accuracy is worth to be investigated. During the stereovalidation procedure of the innovative STC imaging system, several image pairs of an anorthosite sample and a modelled piece of concrete have been acquired under different illumination angles. This set of images has been used to test the effects of the compression algorithm (Langevin and Forni, 2000) on the accuracy of the DTM produced by dense image matching. Different configurations taking in account at the same time both the illumination of the surface and the compression ratio, have been considered. The accuracy of the DTMs is evaluated by comparison with a high resolution laser-scan acquisition of the same targets. The error assessment includes also an analysis on the image plane indicating the influence of the compression procedure on the image measurements.