Physical-Based Automatic Target Recognition Using Angular-Diversity Electromagnetic Scattering Data

A physical-based automatic target recognition (ATR) technique is developed by using angular-diversity electromagnetic scattering data and principal components analysis (PCA). The angular-diversity scattering fields from different concerned structures are pre-calculated to compose the training vectors of PCA. Then, a low-dimensional eigenspace is utilized to efficiently identify the similarity of the obstacles under test. Some typical numerical results show the stable performance of PCA applied to ATR.