摘要
利用目标对对空警戒雷达发射波形的调制效应,采用2种不同的方法,提取出低分辨雷达飞机目标机型(大、小)和飞机目标架次可资分类的特征参数,作为飞机目标机型(大、小)和架次判别的特征向量、特征矢量,然后采用神经网络对目标进行分类识别,给出用BP神经网络进行训练和识别的结果,并在低分辨雷达目标识别样机系统对飞机目标进行分类识别试验中,验证了所提取特征的有效性。表明该方法是有效的,这为常规低分辨雷达空中目标识别提供了一种新的途径。
This paper adopts two methods to abstract low resolution conventional radar and feature parameter of aircraft size and number. It is used as vector,then a neural network is used to classify targets with feature that have been extracted,the results of training and recognition by using BP neural network are given. In the experiment,the effeciency of feature extraction is proved. The experiment result indicates that these methods are valid.
出处
《现代电子技术》
2005年第19期17-19,22,共4页
Modern Electronics Technique
关键词
特征提取
目标识别
熵
神经网络
feature extraction
target recognition
entropy
neural network
