The Statistical Wide-Band Feature Method of HRRP Based on Optimal Path Means

Abstract

High Resolution Range Profile (HRRP) reflects the size and partial structure feature, and is an effective feature for target recognition. A statistic wide-band feature method based on Optimal Path is proposed, and it can extract accurately the statistic feature of HRRP to realize the stable and unstable targets. The method firstly estimates the scattering points for HRRP based on Optimal Path Means, and more calculates the self-relation feature of all scattering points, and utilizes the value of feature to judge the wave characteristic of target movement, finally can recognize stable targets. Simulation examples show that the proposed feature is steady compared with the common wide-band features, such as radial length, central moments feature, and the entropy of HRRP, and can get a well classified result in the low Signal-to-Noise (SNR) condition.