Object-oriented classification of polarimetric SAR imagery based on Statistical Region Merging and Support Vector Machine

Abstract

This paper presents a new object-oriented classification method based on statistical region merging (SRM) for segmentation and support vector machine (SVM) for classification where polarimetric synthetic aperture radar (PolSAR) data are used. The proposed approach makes use of polarimetric information of PolSAR data, and takes advantage of SRM and SVM. The SRM segmentation method not only considers spectral, shape, scale information, but also has the ability to cope with significant noise corruption, handle occlusions. The SVM used for classification takes its advantages of solving sparse sampling, non-linear, high-dimensional, and global optimum problems comparing with other classifiers. It is thus expected that the input vectors of SVM will include fully polarimetric information for image classification. A test image, acquired by the Jet Propulsion Laboratory Airborne SAR (AIRSAR) system, is used to demonstrate the advantages of the proposed method. It is shown that the proposed approach outperforms the traditional pixel-based SVM classification method for land cover classification with PolSAR data, and the integration of SRM and SVM makes the proposed algorithm an attractive and alternative method for polarimetric SAR classification.