%0 Journal Article
%T 一种基于RVM和DS的一维距离像融合识别方法
%A 李睿
%A 王晓丹
%A 蕾蕾
%A 赵振冲
%J 智能系统学报
%D 2016
%R 10.11992/tis.201511021
%X 从如何进一步提升融合识别性能出发，研究有效的高分辨距离像（high range resolution profile，HRRP）融合识别方法。提取了3种平移不变特征，构建了高性能相关向量机（relevance vector machine，RVM）进行特征分类，用DS证据理论融合分类结果以得到目标识别结果，从而提出一种基于RVM和DS的一维距离像融合识别方法。该方法充分利用了RVM输出的概率信息，解决了用DS证据理论进行融合时基本概率赋值获取困难的问题，仿真实验结果表明了本文方法的有效性。</br>Aimed at improving target fusion recognition performance, an efficient approach to radar high resolution range profile (HRRP) fusion recognition is investigated. Three translation-invariant features were extracted from the HRRPs. Meanwhile, a high performance RVM (relevance vector machine) classifier was constructed and DS evidence theory used to fuse the recognition result. A HRRP classification approach, combining RVM and DS evidence theory, is then presented. The method makes full use of RVM output probability information, which solved the difficulty of getting BPA in DS evidence theory. The experimental results based on the simulated data show the effectiveness of the proposed approach
%K 目标识别
%K 一维距离像
%K 相关向量机
%K 证据理论&lt
%K /br&gt
%K target recognition
%K HRRP
%K RVM
%K DS
%U http://tis.hrbeu.edu.cn/oa/darticle.aspx?type=view&id=20160417