Research on Automatic Segmentation of Remote Sensing Image

For the characteristics of remote sensing image, a level set segmentation method using global optimization C-V model based on wavelet multi-scale analysis is proposed to deal with remote sensing image in this article. Wavelet multi-scale analysis is introduced to the level set method for solving slow curve evolution speed. Local average standard deviation is proposed to choose the image optimal scale. Segmentation experiment in RS image has been done in this article, and the result proved to be good. The result show that: The initial contour can obtained best at optimal scale, level set curve evolution efficiency has been greatly improved. The method in this article has optimized the speed and accuracy of image segmentation, it has some practical value.