Abstract: In the ground target recognition, the target echo contains a lot of clutter inevitably, which will seriously affect the recognition performance. To solve this problem, a robust clutter recognition method is proposed in this paper. Firstly, an anomaly detection method based on spherical hypothesis clustering is used to suppress small cluster clutter. Then, the parameter independent local outlier factor (LOF) anomaly detection method is used to detect the residual scattered points and suppress the point clutter. Finally, feature extraction and target recognition are carried out using the high resolution range profile (HRRP) after clutter suppression. The method in this paper has no requirement on whether the target is moving or not, and the parameters involved in the algorithm do not need to be set manually. The experiment results of real measured data show that the proposed method is robust against clutter and the computation is small.

Key words: target recognition, clutter suppression, high resolution range profile （HRRP）, anomaly detection