A NEW IMAGE MATCHING ALGORITHM WITH MODIFIED GREEDY ALGORITHM FOR REMOTE SENSING IMAGERY

摘要

This paper proposes a new approach to image matching by epipolar constraint and local reliability constraint for Remote Sensing Image. We define a new measure of matching support according to the local reliability constraint. A new search strategy—modified greedy algorithm is developed for robust image matching. This strategy only selects those matches having both high matching support and low matching ambiguity. The proposed algorithm has been tested and works remarkably well in remote sensing imagery stereo pairs.