Information Extraction And SVM Classification Of Target On Polarimetric SAR Images

Polarimetric synthetic aperture radar(Pol SAR) can utilize fine structures of different polarization channels which distinguishes SAR complex image of the object, the target point, in remote sensing fields with a wide range of applications. Information extraction from polarization SAR images on the target feature specificity. Pol SAR imaging solutions and applications, has important theoretical and practical value. Paper extract polarization SAR gives information about the main objectives of the study. Focus on polarimetric target decomposition and image classification is the foundation of Pol SAR study.Firstly, the main objectives is to study polarization characteristics and target decomposition method. including target coherent scattering matrix factorization(decomposition with Pauli and SDH test), non-coherent target decomposition on eigenvalues and incoherent target scattering model decomposition method, whichin this paper uses Freeman and Yamaguchi decomposition test. Based on the research polarimetric target decomposition methods and applications, aimed at the man-made features unique scattering properties and structure.By utilizing EMISAR of Pol SAR, data were gathered by studying target decomposition method effectiveness. Artificial information objectives of the scattering components decomposition as the main feature for the decomposition and Pol SAR image classification.Secondly, Pol SAR image classification using SVM(support vector machine) in the case of small sample good properties. This paper studied the method of SVM classification Pol SAR image experiments. EMISAR classification test based on data and analysis of test results and comparation. The proposed algorithm can better detect Pol SAR building man-made objects in the image, with some applicability.