Systems Engineering and Electronics

Previous Articles     Next Articles

Novel supervised classification approach for multifrequency polarimetric SAR data

Biao You1, Bin Xu1, Jian Yang1, Chunmao Yeh2, and Jianshe Song3   

  1. 1. Department of Electronic Engineering, Tsinghua University, Beijing 100084, China;
    2. Beijing Institute of Radio Measurement, Beijing 100854, China;
    3. Xi’an Research Institute of Hi-Technology, Xi’an 710025, China
  • Online:2015-12-25 Published:2010-01-03

Abstract:

A novel method is proposed for the supervised classificationof multifrequency polarimetric synthetic aperture radar(PolSAR) images. The coherency matrices in P-, L-, and C-bandsare mapped onto a 9×9 matrix Ω based on the eigenvalue decompositionof the coherency matrix of each band. A boxcar filteris then performed on the matrix Ω. The filtered data are put intoa complex Wishart classifier. Finally, the effectiveness of the proposedmethod is demonstrated with JPL/AIRSAR multifrequencyPolSAR data acquired over the Flevoland area.