Segmentation optimization simulation of water remote congestion image of the ship

In the process of segmenting water remote congestion image of the ship, due to interference of the external environment and the influence of the number of ships, the target is blocked, traditional methods for image segmentation region intersection, lead to target obscured, therefore, in this paper, a segmentation optimization method for water remote congestion image of the ship based on MCMC is proposed, through morphology denoising to preprocess water remote congestion image of the ship, and remove the noise of the image, ensure do not produce the global geometric distortion. Water remote ship congestion image field is divided into many disjoint areas, to give all states of the image, by using Bayesian method deduce the solution space of state, and setting up four classes of water remote ship congestion image model for four of the most frequent image regions appeared, the solution space structure is analyzed, and data driven method is used to classify the characteristics, according to the probability of each pixel eigenvector belonging to the cluster center to calculate the proposal probability, and transfer speed of Markov chain, establish ergodic Markov chain solution space, so as to achieve segmentation optimization of water remote ship congestion image. The simulation results show that the proposed method not only has the very high accuracy of image segmentation, also has complete segmentation result of target.