Within MAS Framework,Study Of Remote Sensing Image Segmentation

In remote sensing image,the mass and heterogeneity is the emphasis and difficulty of the segmentation.For this purpose,the multi-agent framework is put forward based on statistics and clustering of two models of remote sensing image segmentation methods.The multi-agent system(MAS)is composed of a series of segmentation agents and a cooperation agent.In the method based on statistical model,first of all,assume that pixels in homogenous area are modeled with Gamma distribution with parameters dependent on the type of the region;In order to depict the SAR image pixels in the class attribute,the individual labels form a so-called label field,in which labels are modeled with a prior Potts model where labels of neighboring pixels will tend to be similar;By Bayesian theory,the segmentation model can be built by producing the image model and the label model.To simulate the segmentation model,an approach combining expectation maximization algorithm(EM)and genetic algorithms(GA)within a multi-agent system framework is proposed.Each segmentation agent initializes a global segmentation by EM algorithm,and the cooperation agent employs GA to implement global optimal segmentation.In the method based on clustering model,first of all,an image domain is partitioned into several sub-regions by regular tessellations,and then each segmentation agent algorithm initializes the regional segmentation using Fuzzy C-means(FCM)to realize segmentation of color remote sensing image,within multi-agent framework,each segmentation agent cooperates with other neighboring agents to automatically determine the number of classes in each sub-region,cooperation agent coordinates the regional class attribute and class belongs to the center.In order to verify the feasibility and effectiveness of the proposed method,the segmentation is carried out on the simulation image Synthetic Aperture Radar(SAR)images and color remote sensing image.The experimental results show that the proposed method can solve remote sensing image segmentation.Furthermore,the results of qualitative and quantitative analysis show that the proposed method is effective and promising.