Remote sensing image matching via Harris detector and sift discriptor

Image matching is a fundamental task in the analysis of the remote sensing data. The precision of the image matching have played an important role in the application of remote sensing images, such as three-dimensional modeling, pattern reorganization, panoramic reconstruction, super resolution, phasic analysis, etc. At present, the development of image matching technology and later use in China and abroad is mainly focused on the robust key point, which means that the extracted feature should been invariant to the translation, scale and rotation etc. In recent years, the sift algorithm has been the most popular for extracting interest points. But the calculation procedure is too slow. In this paper, our approach starts with a matching algorithm that combine Harris detector together with sift descriptor method. Our approaches start with the procedure of counting Harris method on the remote sensing image and discard the key point with low contrast. Moreover, we perform an accurate key point location. At last, we use SIFT descriptor as a feature to perform matching from the descriptor point vector in the reference image to the sets of point vector in the querying image.