Research On Target's Feature Extraction From High-Resolution SAR Image

As a tool of acquiring information, SAR is of significant strategic sense, the interpretation of SAR imagery is a very important task. Compared with the capability of acquiring the dala, the processing technology of SAR imagery has dropped it behind and can't meet the requirement of fast process to a great deal of data, so it need to be improved imminently.The feature extraction from SAR image is an important step in the interpretation based on SAR image, it can surpress dimension of the feature space to interpret SAR imagery accurately and fast. In this thesis, we classify SAR targets into point targets, liner targets and region targets roughly via the analysis of SAR imagery, and classify the imagery features into the point features, liner features and region features. These features are extracted from SAR imagery respectively. .Based on the further analysis to the targets' characteristic, the region features are extracted from SAR image at first and two different segmentation methods for SAR image are researched, then we extract the geometrical features and statistical features of the target region segmented. Finally, a modified target azimuth estimation method is given based on the deficit analysis of the existent methods.Edge extraction is not only the basis of the interpretation of liner targets in SAR image, but also can improve the performance of image segmentation. Because the classic edge detector can't be applied to SAR image directly, the RoA algorithm adapting to SAR image is analyzed.In the end, for the problem of extracting the point features from SAR image, peak and ridge are researched chiefly, because they are more significant for SAR imagery. In the thesis, a more reasonable peak model is gained, and the detecting method of peak is designed, then the location stability of peak is analyzed. Besides these, we also extract ridges from SAR image.