Remote Sensing Image Matching Algorithm Based on Harris and Sift Transform

Image matching is a key part of many remote sensing image processing and image analysis. The traditional gray correlation matching algorithm based on corner point because they do not have the rotational invariance requires manual intervention to roughly match can not be automated. SIFT (Scale invariant feature transform) algorithm to solve the image rotation, scaling and other issues, but for the geometry characteristics clearer, richer texture information in terms of the high-resolution remote sensing images, the algorithm consumes more memory, speed of operation is slow the problem is very prominent. The combinations of two proposed image matching algorithms are based on Harris corner and SIFT descriptor. The experimental results show that, compared to the SIFT algorithm, this algorithm significantly cut computation time, while preserving the rotational invariance of the SIFT descriptor and adaptation to light gray correlation algorithm can not overcome disadvantage of fully automatic, in the high better resolution remote sensing image matching.