A Novel Unmanned Aerial Vehicle Recognition Method Based on Bayesian Coherent Point Drift with Full Polarization Information

The utilization of polarimetric information can expand the understanding of UAVs which effectively interprets shape and structure. In this paper, based on the theory of Bayesian Formulation of Coherent Point Drift (BCPD), a novel radar target recognition algorithm has been proposed. Firstly, the canonical scattering centers are extracted based on the radar imaging results of UAVs which helps to understand the structures and scattering properties. Then we build a novel similarity measurement between the model UAV sample and the test samples which not only reflects the prominent scattering centers but also preserve the overall structure features. UAV electromagnetic data can be acquired which is measured in the anechoic chamber or simulated by electromagnetic software. Finally, we fuse three polarization channels with Bayesian criteria to improve the accuracy of recognition results. The experiment proves that the recognition rate of proposed algorithm with full polarization is better than that with single polarization channel.