The Research Of Change Detection Algorithms Based On SAR Images

The change detection of remote sensing images is to quantitatively analyze two remote sensing images of different time segments in the same region,so as to obtain the change information between the two remote sensing images.Synthetic Aperture Radar(SAR)is an active remote sensing technology with the advantages of omnidirectional,high-resolution,large-area coverage.It can penetrate the cover recognition camouflage to accurately obtain ground change information.Therefore,it can be widely used in military survey and reconnaissance,land and resources monitoring,and urban planning.However,SAR image change detection technology still faces many challenges in current practical applications,such as: how to reduce SAR image noise;how to accurately obtain difference images of SAR images at different time in the same region;how to correctly realize classification of difference images;How to more accurately evaluate the performance of the change detection algorithm.Therefore,studying the SAR image change detection algorithm is of great significance to people's real life.The purpose of this paper is to further improve the SAR image change detection algorithm and improve the accuracy of the test results.The research focuses on how to reduce the original noise of SAR images,how to obtain different information,how to accurately classify the detection results and how to evaluate the performance of the algorithm.And achieved the following research results:1.A semi-implicit denoising SAR image change detection algorithm based on ROF model is proposed.Firstly,we use logarithmic transform to transform multiplicative speckle noise into additive noise in SAR image;secondly,we use semi-implicit difference format to solve ROF model;secondly,we use difference method to get difference image from two remote sensing images;finally,the FCM clustering algorithm based on fuzzy C-means is used to cluster the difference images acquired by the difference method,and the final change detection results are classified into change and unchanged categories.The experimental results show that the proposed algorithm further improves the detection efficiency of SAR image change detection and balances the detection accuracy and running time.2.A new SAR image change detection method based on semi-implicit denoising of ROF model and PCA fusion is proposed.Firstly,the semi-implicit difference scheme is used to solve ROF model;secondly,the logarithmic ratio method and neighborhood ratio method are used to obtain the difference image;then,the PCA fusion algorithm is used to fuse the difference image of logarithmic ratio difference map and mean ratio difference map to get the final difference image after fusion;finally,the fuzzy local information C-means clustering FLICM is used to cluster the final PCA-fused difference images,and the final change detection results are changed and unchanged.The experimental results show that the proposed algorithm improves the detection accuracy and reduces the overall running time.