摘要
针对雷达一维距离像样本对样本姿态角的敏感性、样本分布的非线性、样本残缺等问题,提出一种基于非线性零空间雷达目标一维距离像识别方法,通过引入核函数来处理一维距离像目标样本分布中的非线性问题,采用零空间提取最具辨别力的特征信息,最后使用支持向量机处理样本残缺问题并进行分类识别。该方法能有效解决上述问题,并最终提高雷达一维距离像的识别性能。选取3种不同飞机目标的一维距离像进行了仿真实验,识别结果表明了该方法有较高的识别正确率以及良好的抗噪性能。
In the radar target＇s 1-D range profile,the sample is highly sensitive to the aspect variation,the sample distribution is nonlinear,and the sample is often incomplete.The algorithm based on nonlinear null space was proposed in this paper.In this method,kernel method was used for solving nonlinear distribution of radar target＇s 1-D range profile,and null space was adopted to extract the most discriminated characteristic information.At the end,defective samples were handled by Support Vector Machine（SVM）,so does the target＇s classification.The matter mentioned above can be solved by the method,and the accuracy of 1-D range profile＇s classification is improved.Recognition results on three different measured air-plane 1-D range profiles show the competitive classification and noise immunity performance of the proposed method.
出处
《计算机应用》
CSCD
北大核心
2011年第A01期55-57,共3页
journal of Computer Applications
关键词
支持向量机
非线性
零空间
一维距离像
目标识别
Support Vector Machine（SVM）
nonlinear
null space
1-D range profile
target identification
