基于AdaBoost的雷达目标HRRP识别

Radar Target HRRP Recognition Based on AdaBoost

针对雷达高分辨一维距离像(HRRP)特征维数高的特点,采用线性判别分析(LDA)和核Fisher判别分析(KFD)方法进行特征压缩和提取。分析了基于AdaBoost算法的分类器的设计思想和实现步骤。构造了高斯型弱分类器,利用AdaBoost算法集成高斯弱分类器实现了一强分类器,利用此分类器对降维后的HRRP数据进行分类识别,并同K近邻分类器(KNN)、支持向量机(SVM)分类器进行比较,得到一些有价值的结论。

The dimension of the radar high resolution range profile(HRRP) is always high,so HRRP features are extracted by using the linear discriminant analysis(LDA) and the kernel Fisher discriminant(KFD).To obtain an efficient classifier,the AdaBoost algorithm is adopted and a classifier is constructed by using the Gauss weak-classifier.The features obtained from LDA and KFD are inputs of the classifier.Another two classifiers based on the K-nearest neighbour(KNN) and the support vector machine(SVM) are compared wi...