A New Range Profile Identification Method Based on Multi-Scale Entropy Feature

In high resolution radar,one dimensional range profile supplies detailed information of targets,but it is sensitive to direction, translation and intensity.To get direction, translation and intensity invariable features,a wavelet based multi-scale entropy feature is proposed.As entropy is not dependent on the shift of range profiles and is insensitive to the direction variation,the entropy feature has a translation and direction invariable property. In the paper,the entropies of neighbor two scales were divided to eliminate the intensity sensitivity of profiles.When the proposed feature was applied into real data for target identification,good result has been obtained, which is much better than conventional methods. Therefore,the proposed multi-scale entropy feature is an effective feature for range profile identification.