Local Semantic Feature-Based Hierarchical Matching Algorithm for Wide-Area Remote Sensing Images

Remote sensing image matching is an important branch of image matching in computer vision, and with the increasing variety of remote sensing images, the quality differences are also increasing. Aiming at the bottleneck problem of being unable to intelligently match wide-area remote sensing images with significant differences in scale information such as size and physical resolution, this paper proposes a remote sensing image hierarchical matching algorithm based on local semantic features, which solves the problem of low accuracy of remote sensing image matching in the case of large disparity in field of view. The experimental results show that the matching accuracy of the proposed regional hierarchical matching algorithm can be effectively improved compared with the direct matching algorithm of remote sensing images.