Improving Concealed Target Detection in Forest Area Using Fusion of Polarimetric SAR Data and Hyperspectral Data Using Polarimetric and Spectral Signatures

The main purpose of this study is the fusion of polarimetric SAR (PolSAR) data and hyperspectral data for improving concealed target detection in forest area. The proposed method based on the polarimetric signatures, the spectral signatures and the pattern recognition matching methods. The combination of SCM, that measures the similarity between the shapes of two signatures, and SAM, that measures the difference between the intensities of signatures were selected for this study. The results of the proposed method were evaluated using two methods. First, the output of the proposed detector using a fusion of the two source data was compared to the output of the proposed detector using each of the two sources in detection of concealed targets. The number of false detected targets in these methods are 3, 9 and 12. Second, the algorithm was compared to ACE algorithm inhyperspectral data and CFAR method in PolSAR data using the ROC curves. The area under the curve of these methods are 0.87, 0.91 and 0.69. As the results, the proposed method has detected concealed targets with the minimum false targets. Also, this method presented well accuracy in other targets.