Remote sensing image segmentation based on game theory

According to the characteristics of the remote sensing image, the problem of segmentation can be converted to the problem of symbolizing, and finally converted to the solution of Maximum A Posterior (MAP). The global optimum can be found by the algorithm of Simulated Annealing (SA), but it requires a large amount of computations. So sub-optimal algorithm is often used. In this paper we provide a decisive algorithm which is based on 'Game Theory' and prove that it can converge to a local optimum. Because the image we got has too many fragments, we use a grid-algorithm to improve the segmented image and the good result has been got in the experiment.