Target feature extraction based on moment invariants of adaptive time-frequency distribution on chirplet time-frequency atom

Aiming at the problems of feature extraction and recognition of radar target, a target recognition method based on moment invariants of adaptive time-frequency distribution based on chirplet time-frequency atom was proposed. Firstly various moments were analyzed at length, and then change of time-frequency distribution because of various target behavior was analyzed at length by combining radar signal with moment invariant of time-frequency distribution. And then the algorithm was researched at length. At last, the simulation experiment was carried on, the result indicate that the method has higher recognition rate because of fully utilization of target scattering information.