A Method of Residential Areas Extraction in the SAR Based on Texture Features
一种基于纹理的SAR图像居民区提取方法

Residential areas are a familiar object which appears in SAR images,and it has many vital applications in various aspects.A new solution of residential areas extraction is presented in this paper.According high gray value and texture feature of residential areas,we convert the SAR image to discrete points of residential areas with Sobel operator and Otsu threshold algorithm,then we segment it in several blocks with Kmeans clustering,finally implement automatic residential areas extraction using Graham algorithm.Both the theoretical analyses and the experimental results indicate that the proposed method is a highly efficient method of residential areas extraction.