基于多阶段分类的雷达目标识别

讨论了用一种新的识别系统识别高分辨雷达距离像（ＨＲＲＰ），它先在Ｋ－Ｌ变换对距离像进行特征压缩后，用聚类分析进行识别呀基于某个规则排除某此种类，再对不能识别的模式，用种类少时效果好的最佳鉴别准则对模式进行特征提取，然后用模糊ＡＰＴＭＡＰ网络组进行分类，实验表明该方法比常规的单分类器方法有更高的识别率。

RANGE PROFILE IDENTIFICATION BASED ON MULTISTAGE CLASSIFICATION

A novel range profile identification system based on multiple classification was discussed. After feature compression with K L transform, one can get a rough identification result by clustering or eliminate impossible candidates according to some rules. The remainder is identified by fuzzy ARTMAP neural network with the feature extraction of the optimal set based on the Fisher criterion. The experiment shows that this classifying method could be more effective than the single classifier method.