Stereo Matching of Urban Remote Sensing Image Based on Edge Block Coincidence Rate

With the continuous development of remote sensing technology, the acquisition of high-resolution remote sensing image information has become more simpler and faster. Stereo matching, as a basic part of the field of computer vision, stereo matching has a direct impact on post processing. Because the information of remote sensing images is more complicated and the interference is greater than that of ordinary images, it is more difficult to match real remote sensing images. Many stereo matching algorithms are not suitable for remote sensing images. In this paper, a new framework based on the edge block coincidence rate is proposed for the stereo matching problem of urban remote sensing images. The difficulty in stereo matching of remote sensing images lies in the fact that the epipolar lines of remote sensing images captured by linear CCD sensors carried by remote sensing satellites are quadratic curves rather than a straight line, so accurate epipolar line correction images cannot be obtained, thus limiting the matching search space to one dimension. This paper proposes a new stereo matching framework for urban remote sensing image, and experiment results show that the framework can achieve the desired matching effect. In order to solve this problem, this paper proposes to give up the use of epipolar constraint, choose the method of image registration, and adopt the stereo matching method based on mutual information to obtain the final accurate matching result. Experiments show that the framework can achieve ideal matching effect.