Integrating Change Vector Analysis, Post-Classification Comparison, and Object-Oriented Image Analysis for Land Use and Land Cover Change Detection Using RADARSAT-2 Polarimetric SAR Images

Abstract

This study proposes a new method for land use and land cover (LULC) change detection using RADARSAT-2 polarimetric SAR (PolSAR) images. The proposed method combines change vector analysis (CVA) and post-classification analysis (PCC) to detect LULC changes using RADARSAT-2 PolSAR images based on object-oriented image analysis. A hierarchical segmentation was implemented on two RADARSAT-2 PolSAR images acquired at different times to delineate image objects. CVA was applied to the coherency matrix of PolSAR images to identify changed objects, and then PCC was used to determine the type of changes. The classification of the RADARSAT-2 images is based on the integration of polarimetric decomposition, object-oriented image analysis, decision tree algorithms, and support vector machines (SVMs). In comparison with the PCC that is based on the Wishart supervised classification, the proposed method improves the overall error rate for change detection and the overall accuracy for change type determination by 25.15 and 6.59 % respectively. The results show that the proposed method can achieve much higher accuracy for LULC change detection using RADARSAT-2 PolSAR images than the PCC that is based on the Wishart supervised classification.