一种基于调制特征向量的空目标识别方法  

One method of classifying the air-targets based on feature vector modulation

作　　者：苗振奎[1] 徐伟[1] 李亮[1] 张同舟[1] 

机构地区：[1]f江苏自动化研究所,江苏连云港222061

出　　处：《太赫兹科学与电子信息学报》2013年第6期922-926,共5页Journal of Terahertz Science and Electronic Information Technology

基　　金：国防“十一五”基础科研资助项目(62201070701)

摘　　要：空中目标旋转部件对雷达回波有调制作用。将回波时域幅度调制信息的幅度偏差系数作为调制特征,首先利用仿真数据分析了固定翼飞机目标和直升机飞机目标的差异,随后提出了一种最近邻分类器的设计方法,最后利用实测数据对目标进行分类,取得了较高的识别率。与支持向量机的算法相比,该分类器算法复杂度低,计算量小,识别率高,在工程上可以选用这种方法对两类目标进行分类识别。The rotating components of aerial target can have modulation effect on radar echo. Taking the rate deviation coefficient as a modulation characteristic, the differences between fixed-wing aircrafts and helicopter aircrafts are analyzed with simulation data. A nearest neighbor classifier method is proposed. The targets are classified by using the measured data, and a higher recognition rate is obtained. In contrast with the support vector machine algorithm, the proposed method bears low complexity, less computation and high recognition rate. It can be used to classify two types of targets in engineering project.

关 键 词：目标识别 调制特征 幅度偏差系数 最近邻 分类器 

分 类 号：TN957.51[电子电信—信号与信息处理]