Automatic Tie Pointer for <em>In-Situ</em> Pointing Correction
- Monday, 01 August 2011
The program takes the input images and finds overlaps according to the nominal pointing. It then finds the most interesting areas using a scene activity metric. Points with higher scene activity are more likely to correlate successfully in the next step. It then uses correlation techniques to find matching points in the overlapped image. Finally, it performs a series of steps to reduce the number of tie points to a manageable level. These steps incorporate a number of heuristics that have been devised using experience gathered by tie pointing mosaics manually during MER operations. The software makes use of the PIG library as described in “Planetary Image Geometry Library” (NPO-46658), NASA Tech Briefs, Vol. 34, No. 12 (December 2010), page 30, so it is multimission, applicable without change to any in-situ mission supported by PIG.
The MARSAUTOTIE algorithm is automated, so it requires no user intervention. Although at the time of this reporting it has not been done, this program should be suitable for integration into a fully automated mosaic production pipeline.