ECM-based Segmentation For Multi-look SAR Image

The intensities of pixels in multi-look SAR images are usually modeled by Gamma distribution,because of the existence of the Gamma function,EM algorithm not only cannot solve the Gamma function by solving the expectation function of the Gamma function,but also cannot realize the multi-look SAR images segmentation with the variable class.In this thesis,we propose a fixed class segmentation of multi-look SAR imgaes based on ECM algorithm and variable class segmentation of multi-look SAR images combined with BIC criterion and ECM algorithm.Experiments show that the shape parameter can be used as a variable to obtain better segmentation results,and the proposed variable segmentation algorithm can automatically determine the number of SAR images categories.In order to verify the feasibility and effectiveness of the proposed algorithm,the paper proposes a fixed and variable class segmentation experiment for the proposed algorithm,and compares it with the contrast algorithm.The experimental results show that the proposed two methods can realize the SAR images segmentation and the automatic determination of the number of categories,and the qualitative and quantitative evaluation of the segmentation results.The validity and feasibility of the two algorithms are verified by qualitative and quantitative evaluation.