支持向量机的OTHR多频特征目标分类识别法

A Method of OTHR Target Classification and Recognition Based on SVM

针对低信噪比下多频法在天波超视距雷达(Over-The-Horizon Radar,OTHR)目标分类识别中分类精度不高的缺点,充分利用雷达量测的一些先验信息,将基于支持向量机(Support Vector Machine,SVM)的多分类器引入多频法中,提出了一种基于SVM的OTHR多频特征目标分类识别方法。仿真结果表明,利用较少的3个频率点,在信噪比较低的条件下可获得较好的分类识别效果,说明了该方法在OTHR目标分类识别中的有效性和可行性。

The target classification accuracy of multi-frequency method is not high at low SNR in Over-The-Horizon Radar（OTHR）.An algorithm of multi-frequency features target classification and identification for OTHR based on Support Vector Machine（SVM） is given.It combines the multiple classifiers based on SVM with the multi-frequency method,and makes full use of some prior information from radar measurements.Simulation shows the better classification and identification results in the case of using three frequency points at low SNR,which indicates the validity and feasibility to apply this algorithm in OTHR for target classification and recognition.