A modified method for the estimation of SAR target aspect angle based on MRF segmentation

The estimation of a target aspect angle is an important research problem in synthetic aperture radar (SAR) target recognition. Thus, this paper is undertaken to study the algorithm for the estimation of the aspect angle based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) data set. The study employs segmentation based on the Markov Random Field (MRF) to separate the target, shadow, and background. It then proposes two divergence indices, target divergence index (TDI) and joint divergence index (JDI) of target and shadow, to describe the distribution of pixels in target and shadow. The content of the paper is as follows. The first part presents an analysis of the relationship between aspect angle and two indices, and extracts the common features of the aspect angle and the two indices. The second part analyzes the advantages and disadvantages of the three methods. The last part presents the proposal for a combined algorithm for tank aspect angle estimation. The results show that the proposed algorithm is effective.