Aimed at the problems of single construction,poor noise resistance and insufficient texture features of difference map,a SAR image change detection algorithm based on improved difference operator and Gabor_ELM is proposed.Combining the efficiency of the improved logarithmic ratio and the robustness of the neighborhood ratio,the method proposed an improved difference operator,and extracted Gabor texture information through the difference map combined with hierarchical fuzzy clustering and extreme learning machine(ELM)to achieve Gabor_ELM change detection algorithm.Experiments show that compared with FLICM,MRFFCM,ELM and PCANet algorithms,the proposed algorithm can better retain the change details and has higher detection accuracy.The correct rate of the experimental data is over 98％,and the Kappa coefficient of the Ottawa data reaches 0.94.
维普
万方数据