基于模型融合的窄带雷达空中目标分类方法  

An air target classification method based on model fusion for narrowband radar

作　　者：吴鑫 孟凡君 WU Xin;MENG Fan-Jun(No.8 Research Academy of CSSC,Nanjing 211153)

机构地区：[1]中国船舶集团有限公司第八研究院,南京211153

出　　处：《雷达与对抗》2022年第4期15-19,共5页Radar & ECM

摘　　要：针对窄带雷达3类飞机分类问题,提出一种基于模型融合的窄带雷达空中目标分类方法,在多个特征域对目标进行特征提取,通过Stacking策略对K近邻、随机森林和支持向量机进行模型融合,利用融合模型实现3类空目标的分类,最后通过测试验证了其有效性。Aiming at the classification problem of three types of aircraft for narrowband radars,a classification method of air targets is proposed based on the model fusion.The target features are extracted in multiple feature domains,and the model fusion is performed on the K-nearest neighbor,random forest and support vector machine through the Stacking strategy.The classification of three types of air targets is realized by using the fusion model.Finally,the test data show that the method is effective.

关 键 词：目标分类 窄带雷达 模型融合 特征提取 

分 类 号：TN957.52[电子电信—信号与信息处理]