An Information Geometry Detector for Space-time Adaptive Detection Based on KL Divergence

Slowly moving and weak target detection is a serious challenge for airborne radars, especially in heterogeneous clutter environment. For the reason that the height of airborne radar is high, the airborne radar usually faces a much more heterogeneous clutter environment and requires better anti-clutter ability for the detector. To address this problem, a novel KL-divergence-based matrix information geometry detector for space-time adaptive detection is proposed . The proposed method first transform the target detection problem into matrix manifold space, and then estimate the clutter covariance matrix with the manifold distribution characteristic of reference samples. By processing the radar echo in the nonlinear manifold space, the proposed method achieves better clutter suppression ability and target detection ability. The experimental results based on the Sea-detecting Data-sharing programe data validate the superior performance of the proposed algorithem.