New Classification Method Based on Cloude-Pottier Eigenvalue/Eigenvec torDecomposition

摘要

In this paper, a new polarimetric scattering parameter, the averaged intensity (I), is introduced to present the backscatter intensity of fully polarimetric Synthetic Aperture Radar (SAR) data. According to the particular analysis on the properties of I, a angle, and entropy H, the mapping rule of I-a-H feature space onto the Intensity-Hue-Saturation (I-H-S) color space is proposed. The authors use the IHS transform instead of a segmentation algorithm to finish the classification. The important advantages of this method are that the information contained in the I-a-H feature space is preserved without any loss in the resulting image, and the execution time is saved since the IHS transform is faster than most complicated segmentations. The result shows that I has the additional information that is not contained in a and H, and the classification result is more readable than that of H/a/A (entropy, scattering angle, anisotropy).