Adaptive Targets Detection Of Wideband Polarimetric Radar

The high-resolution polarimetric radar can obtain more information of the target characteristic than conventional ones. Exploiting these polarization characteristics sufficiently can remarkably improve the detection and identification performance. Some research results and application experience of high-resolution polarimetric radar have been made. Therefore, research on the adaptive polarization detection algorithm of the distance distribution target is of great significance value.The polarimetric radar can obtain the entire information content of the received wave by vector measurement process, while conventional radar systems cannot obtain the polarization characteristics of target and clutter. Firstly, the essential tools for radar polarization analysis are introduced, and the essential polarization behavior of land and sea clutter is described in this paper.In fact, the Gaussian assumption can be no longer met for radar clutter in high- resolution radars (HRRs) and/or at low grazing angles. More specifically, the disturbance can be better described as a compound-Gaussian process. The spherically invariant random processes (SIRP) model is introduced, and the simulation method of K distribution is described. In this paper, a polarimetric adaptive algorithm, based on Generalized Likelihood Ratio Test (GLRT) criterion, for the detection of range-distributed targets against compound-Gaussian clutter background is presented. The theoretical expression of probability of false alarm (PFA) is derived. The proposed detector ensures the theoretically constant false alarm rate (CFAR) respect to the texture statistics and the clutter covariance matrix.Finally, we analyze the performance of the algorithm based upon simulated clutter. The simulation result show that the detector presented in this paper achieves a better detection performance than its counterparts with a conditional signal to clutter ratio (SCR). And the false alarm stability of the proposed detector is tested and verified by simulation.