一种窄带雷达快速目标分类方法  

Fast target classification method for narrowband radar

作　　者：董会方 卞磊[1] 杨广平[1] Dong Huifang;Bian Lei;Yang Guangping(Nanjing Institute of Electronic Technology, Nanjing 210013, China)

机构地区：[1]南京电子技术研究所

出　　处：《国外电子测量技术》2018年第2期105-111,共7页Foreign Electronic Measurement Technology

摘　　要：近年来,窄带雷达目标识别技术的理论研究已日趋成熟,但由于受限于雷达自身工作方式、缺乏完整的数据库以及信息的实时处理能力,其工程应用方面还处于初期阶段。针对工程应用中提高识别率必须以牺牲雷达时间资源这一问题,文章研究了在窄带雷达体制下的目标分类问题,先利用航迹特征对飞机类目标、巡航导弹类和制导炸弹类3类目标进行粗分类;针对飞机类目标,再基于JEM的调制特征区分直升机和螺旋桨飞机,进行细分类;采用SVM（支持向量机）分类算法;分类方案采用人工识别和自动识别相结合,人在回路的判决策略;在雷达的正常工作模式下,通过实测数据分析,取得了较好的识别率。In recent years,the theoretical research on narrowband radar target recognition technology has become more mature,but because of its limited working mode,lack of complete database and real-time processing of information,its engineering application is still in its early stages.In order to improve the recognition rate in engineering applications,we must study the problem of target classification under the narrow-band radar system.At the same time,we study the target classification problem in the narrow-band radar system,and use the track characteristics to target the aircraft,cruise missile and guided bomb.And SVM（support vector machine）classification algorithm is used to classify the helicopter and propeller aircraft based on the modulation characteristics of JEM.The classification scheme adopts the combination of artificial identification and automatic identification.Decision strategy;in the radar normal mode of operation,through the measured data analysis,and achieved a better recognition rate.

关 键 词：窄带雷达 目标分类 粗分类 人工识别 自动识别 

分 类 号：TN957.51[电子电信—信号与信息处理]