CFAR Method Based On Homogeneity Similarity For SAR Image Change Detection

Synthetic aperture radar (SAR) system has an all-weather, all-time, high resolution and strong penetrating ability, therefore, the SAR image has been widely applied in national security and national economy. Change detection of remote sensing images means to detect the change information by analyzing the remote sensing images which is acquired at the same area with different time. It is one of the important applications of remote sensing technology.With the development of remote sensing technology and image information processing analysis technology, the research on change detection of remote sensing images becomes an important research direction about remote sensing image, and widely used in social and economic fields, such as land resources monitoring, water monitoring, environmental monitoring, urban management planning, forestry monitoring, agricultural research, disaster monitoring and evaluation,vegetation coverage and meteorological monitoring and so on.This paper revolves around how to effectively build difference image and effectively detect change information between two temporal remote sensing images, which completes the following three aspects, the main work includes the following three aspects:(1) A algorithm CFAR threshold SAR image change detection based on block similar statistical distribution is presented. The algorithm designs ratio method considering the multiplicative noise of SAR image, to get the distribution of ratio probability. The method can solve the problem of the relationship between threshold and false alarm. Compared with the existing point-based method, neighborhood information is considered and replaces the point ratio with block means ratio to reduce the noise interference and get more stable detection results.(2) A refined CFAR threshold SAR image change detection method is presented. The method is based on the shape adaptive patch (SAP), which combining homogeneity similarity and structure similarity of image, it perform better in change detection of point target, edge line target, and the accuracy is higher.This work was supported by the Fund for Foreign Scholars in University Research and Teaching Programs (the111Project)(No. B07048), and the Program for Cheung Kong Scholars and Innovative Research Team in University(No. IRT1170).