Performance Experiment of Classification Using Chinese Airborne Multi-Band and Multi-Polar SAR Data

Multi-band SAR and multi-polar SAR is one of the most important directions for SAR system, and with the developments of multi-band and multi-polar SAR, it is of interest to compare the efficiency of classification using multi band and multi-polar SAR images. In this paper the classification was experimented with the airborne multi-band and multi-polar SAR data which were acquired in April, 2010, by East China Research Institute of Electronic Engineering, China Electronics Technology Group Corporation (CETC), over the test site near Sanya city, Hainan Province, China. It is analyzed that the performance of classification of each channel and the fusion image, and given some conclusions at last. Before the classification are drawn experiment, remote sensing data must be preprocessed, including calibration, filtering and image fusion, with the ROI defined. Then, ENVI software was used on experimental image classification. Through analysis of the classification accuracy of the results, it is found that the fused image based on multi-band overall classification accuracy of the 99.2195%, that than single-band image accuracy is increased by more than 14%, and Kappa coefficient from 77% to 98.82%.