Polarimetric SAR Image Building Extraction Method

The classification, recognition and extraction with polarization SAR data are important issues in SAR image interpretation, and have much theoretical and applicable significance. The extraction of POLSAR image is an important step for image interpretation, and the extraction results can provide auxiliary information for edge detection, target detection and recognition, and can be as the final results to provide service for users. Therefore, this dissertation gives a deep research in two aspects of the scattering mechanism and polarization characteristic information, and the purpose is to extract buildings.This dissertation firstly introduces analyzing SAR image classification current research situation at home and abroad, then introduces the polarization SAR basic theory and knowledge and the relevant knowledge of polarization target decomposition. Then using the classification method based on scattering mechanism to extract buildings, through the analysis of the results, points out that the simple classification method based on scattering mechanism in building extraction lack of completeness and accuracy.Then we take a deep study of the buildings classification of polarization SAR image. First analyzing the accuracy and integrity of the traditional methods which using scattering mechanism to extract buildings, and then we combine the polarization characteristic parameters to extract buildings, so this dissertation proposes a new method which based on scattering mechanism and polarization characteristic parameters to extract buildings. Then discusses the method and polarization of the scattering mechanism of three different buildings combined with the characteristic parameter extraction method, presents the corresponding building extraction of experimental results, and made a detailed analysis and comparison with the result of the experiment. Results show that the new method can make up for the traditional methods due to insufficient information extraction results inaccurate and incomplete.Finally summarizes the main research work and innovation of the dissertation, and puts forward the further research plan.