Research And Application Of Image Segmentation Algorithms In SAR Images

Synthetic Aperture Radar is an active microwave remote sensing, which is an important branch in the field of remote sensing. Synthetic Aperture Radar has the characteristics of high resolution, strong penetrating, all-time, all-weather and so on. It can be used in many fields. With the rapid development of SAR system, the number of SAR images is getting more and more, but the ability to understand and interpret SAR images is relatively lagging.SAR image segmentation is one of the key steps in SAR image interpretation, and it is the basis and premise of SAR image understanding and interpretation, and has been widely concerned by domestic and foreign scholars. However, due to the inherent imaging mechanism of SAR image, speckle noise is present in the SAR image, which makes the previous image segmentation algorithm can't obtain satisfactory results in SAR images.Therefore, it is very important to study the segmentation algorithm of SAR image.This paper analyses and studies three kinds of SAR image segmentation algorithms: image segmentation algorithm based on Markov random field, image segmentation algorithm based on fuzzy c-mean and immune genetic algorithm for single threshold segmentation.The segmentation algorithm based on Markov random field distribution is one of the most important models, the basic principle is the use of conditional probability to describe the distribution of image data; the image segmentation algorithm based on fuzzy c-mean is a kind of non-supervised learning, which integrates the fuzzy theory and the clustering,which makes the problem of the analysis more close to the real life; in immune genetic algorithm for single threshold segmentation, the immune clonal selection algorithm is used as the optimization algorithm, which is formed by the clonal selection mechanism of Burnet. In the optimization strategy, the random initialization and continuous generation of new antibodies are used to replace the lower affinity of the antibody, which can ensure the diversity of the population. In this paper, the three algorithms respectively do a simulation experiment, and know the adaptation of each kind of algorithm, and the limitation of the algorithm.