SAR Image Classified Based On Genetic Optimization Neural Networks

It is all-weather detection capability,penetrating strength to be strong looking sideways way Polaroid heavy and close to characteristic such as being real-time area and so on for Synthetic Aperture Radar(SAR) to have,So SAR have already widely employed on the sea oil- pollution monitoring.Most of study has found that oil spill appeared as dark slick or dark spot in oil-spilled SAR image due to the dampening effects on the sea.Utilizing this principle,Oil-spilled accidents that SAR can be used for spotting the sea to happen.The paper makes the SAR oil-spilled images classified,mainly concentrating on the noise reduction algorithms of SAR oil-spilled images,the adaptable characteristics based on GLCM of SAR oil-spilled images and Neural Networks optimized by Genetic Algorithm classifying SAR oil-spilled images.In the research of despeckling in SAR images,the forming mechanism and statistic characteristic of noises are introduced first and then,the dissertation particularly introduces the main noise reduction algorithms.After analyzing the several methods in theory,the several algorithms are tested by using one real SAR oil-spilled image.In order to precisely classify SAR oil-spilled images,the paper raises one method that the adaptable characteristics and Neural Networks optimized by Genetic Algorithm classify SAR oil-spilled images.In the progress of texture analysis,four pixel texture parameters that are sensitive to SAR image of oil-spilled are calculated by gray level difference statistics.The SAR image of oil-spilled is classified by using the feature vector that is composed of the Gray Level Co-occurrence matrix features and gray of pixel.Utilizing Genetic Algorithm optimizing the weights in Neural Networks,the weights are regarded as the initial weight value of the Neural Network Algorithm first and then making use of Kohonen Neural Network to search the best weight value forming a kind of mixing training algorithms to achieve the goal of optimizing the Neural Network.After the whole Kohonen Neural Network builds up and expressed by the expression method of advanced computer language,appraised assignment, reproduction,cross and make a variation,solving by Genetic Algorithm to it space orient some better search space and then continue asking and solve out the categorized result with the Neural Network Algorithm.The results of the above 2 types of network are compared in this paper.All of them have a good performance of classification;but Kohonen Neural Networks optimized by Genetic Algorithm is more effective and accurate one as classifier for SAR images of oil-spilled.