Terrain Classification And Aircraft Target Detection Of PolSAR Image Based On Scattering Characteristics

Polarimetric Synthetic Aperture Radar(PolSAR) is a new kind of coherent microwave remote sensing radar imaging system. It can obtain more plentiful polarimetric information describing the backscattering properties of land cover and land use by transmitting and receiving electromagnetic in different polarization modes. As two major constituents in the interpretation of PolSAR data, image classification and target detection have been playing an important role in the military, agriculture, marine science, geology science and other fields.Studies of two aspects in the interpretation of PolSAR data are focused on in this paper,respectively image terrain classification and aircraft target detection.In terms of image classification, analyzing the terrain classification preserving polarimetric scattering characteristics based on Freeman decomposition and the vantages of three-component scattering model with deorientation, a new unsupervised classification of PolSAR images with scattering model and Wishart classifier is presented. Firstly, the three-component scattering model with deorientation is applied to classify PolSAR images into three fundamental scattering categories and a mixed scattering category. Then, pixels in each kind of fundamental scattering categories are divided into more clusters based on the power of dominant scattering ones. After that, small clusters of each category are merged based on the Wishart distance between clusters. Finally, PolSAR data is classified iteratively with Wishart classifier to achieve unsupervised classification results. Real polarimetric datas are used to verify the new scheme, experimental results show the new algorithm improves classification accuracy and lowers volume scattering overestimation compared with traditional classification.In terms of aircraft target detection, the polarimetric scattering characteristics of aircraft targets is analyzed and a new method of aircraft target detection based on Span Polarization Cross Entropy is proposed. The method calculates the scattering similariy parameters between each pixel and three canonical scattering mechanisms, which are respectively surface scatter, double-bounce scatter and canopy scatter. Then, the scattering similarity parameters are enhanced and normalized to redefine Polarization Cross Entropy(PCE). Span Polarization Cross Entropy(SPCE) is composed with PCE and the scattering span of thetarget pixel. Finally, threshold segmentation is applied for PolSAR data with SPCE and other scattering features to achieve aircraft detection result. Real PolSAR datas are used to verify the effectiveness of the proposed scheme.