Skip to Main Content
In this paper, the authors propose a new method for supervised target classification of polarimetric synthetic aperture radar (SAR) image, by using the optimization of polarimetric contrast enhancement (OPCE). First, using the idea of the generalized optimization of polarimetric contrast enhancement (GOPCE), the authors modify the model with three polarimetric parameters which are related to the physics of the scattering mechanisms. It leads to enlarge the difference between two categories and improve the classification results. A new classification approach is then proposed, it is similar to a single binary tree, which the misclassification between the classes with a big power difference is minimal. After the classified results are obtained by the combination of Fisher-OPCE and polarimetric parameters, the coefficients of the scattering parameters information of every two adjacent classes will be used as the last discrimination for final classification results. The effectiveness of the proposed algorithm is demonstrated by using a NASA/JPL AIRSAR L-band image over San Francisco.