Decision Fusion Of Multi-band SAR Target Detection And Classification Results

Multi-band SAR images provide complementary and redundant information. Based on this characteristic, the target of multi-band information fusion is to get more exact and reliable decision by fusing the multi-band SAR information than only using single-band information. This paper mainly discusses the topic of decision fusion of the processing results of multi-band SAR images, including the target detection algorithm, the classification algorithm and the decision fusion algorithms.There are not only homogenous regions but also nonhomogenous regions in SAR images. To exactly detect the targets not only in homogenous regions but also in nonhomogenous regions, this paper chooses the VI-CFAR detector for multi-band SAR target detection. Based on the detection results, to improve the detection performance of multi-band SAR system, this paper proposes a decision fusion method based on Neyman-Pearson rule to fuse the detection results. Experiment results show that this method can not only get all of the targets and show them in one image exactly, but also can improve the multi-band SAR target detection probability. In the study of multi-band SAR images classification, this paper chooses the classification method based on image segmentation. In this method segmentation uses the special information, so it can restrain the influence of speckle noise to classification. Based on the classification results, to improve the classification veracity of multi-band SAR system, this paper proposes a decision fusion method based on Dempster-Shafer evidence theory to fuse the classification results. Experiment results show that this method can improve the identification rate of multi-band SAR images classification.In the end of this paper, target detection algorithm,classification algorithm and decision fusion algorithms are combined to label the state of the targets.