A Homogenous Reference Cells Selector for Cfar Detector in Highly Heterogeneous Environment

This paper considers the radar scenes which contain numerous rapidly changing terrains, i.e., there are more than one clutter-edge in the environment. This kind of radar scenes incurs sharply degradation in the performance of the present adaptive constant false alarm rate (CFAR) detectors as the statistical characteristic of reference cells is highly heterogeneous. To solve this problem, we propose a homogenous reference cells selector to improve the performance of CFAR detector in highly heterogeneous environment. The selector is comprised of an M -N clutter-edge detector cascading a terrain classifier. The M -N clutter-edge detector is used to obtain multiple clutter-edges in heterogeneous environment. With the detected clutter-edges, the terrain classifier is derived to obtain identical distributed range cells. Based on the selector, a modified Log-t-CFAR detector is suggested. Finally, the performance of the proposed selector and CFAR detector is evaluated by measured data and computer simulation.