Photogrammetric analysis of the Mars Global Surveyor mapping data

Abstract

This paper studies the photogrammetric mapping properties and capabilities of the Mars Global Surveyor (MGS) mapping data. Starting from the raw MGS data, we decompress the MOC narrow angle images, extract, and calculate their exterior orientation from the SPICE kernels, and calculate the 3D coordinates of MOLA footprints from MOLA PEDR files. A new approach is proposed that registers a MOLA profile to stereo MOC images over the same area with robust and faster convergence. Intersection is conducted to determine the 3D positions of image points measured on MOC stereo pairs. It is shown that there is a nearly constant uncertainty of one MOLA ground spacing distance (approximately 325 m) along the flight direction in MOC and MOLA registration. This is caused by the uncertainties in SPICE kernels, MOLA points, and the determination of time tags for MOC scan lines, which possibly constitutes the dominant error source for the registration. Intersection calculation reaches an optimal balance by distributing the uncertainty evenly in the two images of a stereo pair. As for the photogrammetric mapping capabilities, an uncertainty of 180.8 m in planimetric distance and 30.8 m in elevation difference is estimated. A number of numerical and graphic results over three of the selected candidate landing sites for the Mars Exploration Rover mission are presented for analysis and illustration.