Radar HRRP target recognition based on Coherence Reduced Stagewise K-SVD

A novel dictionary learning method, called Coherence Reduced Stagewise K-SVD (CRSK-SVD), is proposed for radar high-resolution range profile (HRRP) target recognition, which is motivated by the desire to overcome the ineffective classification results of the redundant dictionary in the sparse representation based classifier. The proposed method is an adaptation of the popular K-SVD algorithm. It can train a dictionary dynamically by trimming the redundant atoms and adding new efficient atoms according to the sparse representations of the dataset. The experimental results based on simulated radar HRRP targets recognition show that the proposed method can raise the correct recognition rate compared with classical classification methods. Also, the number of atoms is less.