Recognition of Ships and Chaff Clouds Based on Sophisticated Polarimetric Target Decomposition
Sinong QUAN,
Hui FAN,
Dahai DAI,
Wei WANG,
Shunping XIAO,
Xuesong WANG
Affiliations
Sinong QUAN
The State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System, National University of Defense Technology, Changsha 410073, China
Hui FAN
The State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System, National University of Defense Technology, Changsha 410073, China
Dahai DAI
The State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System, National University of Defense Technology, Changsha 410073, China
Wei WANG
College of Electronic Science, National University of Defense Technology, Changsha 410073, China
Shunping XIAO
The State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System, National University of Defense Technology, Changsha 410073, China
Xuesong WANG
The State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System, National University of Defense Technology, Changsha 410073, China
The recognition of ships from chaff cloud jamming is challenging because they have similar dimensions and radar cross sections. In this paper, we propose a polarimetric recognition technique with sophisticated polarimetric target decomposition. Three sophisticated scattering models are integrated to constitute a seven-component model-based decomposition method so as to accurately characterize the dominant and local scattering of ships. Based on the concepts of contrast and suppression, a robust scattering contribution difference feature is designed according to the derived scattering contributions. The constructed feature vector, combined with the polarization scattering angle, is inputted into the support vector machine to fulfill the recognition process. Simulated and real polarimetric radar data are utilized to test the proposed method, and the results show that the proposed method outperforms state-of-the-art methods by achieving the highest recognition rate of over 98%.
