SAR Target Configuration Recognition Based on Integrated Graph Model
Ming Liu, Shichao Chen · 2021
Focusing on the problem of target configuration recognition in synthetic aperture radar (SAR) images, a recognition method using an integrated graph model (IGM) is proposed in this paper. The information of the samples is preserved in different views. To achieve stronger discriminative power, a modified sparse graph employing sign correction of the sparse coefficients is proposed to describe the relationships of the samples. Furthermore, to realize more complete information capturing, the classical Euclidean distance-based affinity matrix together with the proposed modified sparse graph are fused to achieve an integrated graph. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) datasets verify the effectiveness of the proposed method. Satisfying configuration recognition results can be obtained by using the proposed methods.