Recognize Aircraft in ISAR Images

This paper provides a novel method to recognize aircraft in Inverse Synthetic Aperture Radar (ISAR) images. The method utilizes the conspicuous scatterers located in a two-dimensional ISAR image as the feature points of aircraft to generate geometric invariants and recognize the aircraft. When the ISAR imagery is influenced by noise or target rotation, the proposed method provides an effective recognition. To alleviate the computational complexity, the ordered geometric invariants and the primary key are employed. Moreover, dynamic thresholds are applied to increase the accuracy of feature comparisons. The experimental results show that the proposed method is robust and effective to identify aircraft in ISAR images.