A remote sensing image subpixel matching algorithm combined edge with grey

A remote sensing image matching is one of the important and difficult tasks in computer vision and digital photogrammetry. In this paper, it subpixel matching algorithm combined edge with grey is presented in order to match precisely and reliably a pair of remote sensing images taken by dissimilar sensor at different times. The algorithm mainly contains three processing steps. First, edges of objects are extracted with a fast LOG operator. Then, a coarse matching is implemented by using the principle of invariant line moment. Finally, under the constraint of the above coarse matching, a subpixel location will be directly produced with grey-based Least Square Matching (LSM). The experiments demonstrate that the purpose desired is achieved.