基于子像空间的雷达目标一维距离像识别

摘要

A novel approach of radar target recognition is proposed in this paper. First, each target subprofile space formed by training subprofiles of the corresponding target is used as library target template to improve classification accuracy for “known” targets that are included in training target set. Second, according to hypothesis testing theory, the decision threshold is introduced to distance minimum classifier in order to reject effectively “unknown” targets that are not included in training target set. The simulated experimental results on four radar targets show that this method is effective.%提出了一种基于子像空间的雷达目标一维距离像识别方法。该方法一方面通过建立目标子像空间，改善了对输入目标为库属目标的识别率；另一方面基于假设检验理论，在分类判决中引入判决门限，能够有效地对库属目标和非库属目标进行判别。模拟实验结果表明该方法的有效性。