基于二叉树直觉模糊SVM的弹道目标HRRP识别

摘要

弹道目标识别是一个多目标识别过程,二叉树支持向量机(BTSVM)是一种针对多类分类有效的分类器.BTSVM结构简单,训练快,但容易出现误差积累.为提高目标识别率,引入并改进了直觉模糊支持向量机,设计了一种基于自适应隶属度函数和直觉指数的多类直觉模糊SVM分类器.对训练集进行直觉模糊化处理得到直觉模糊数据集,用于训练分类器得到二叉树直觉模糊SVM分类器.将此分类器应用于弹道目标HRRP识别,提高了识别的正确率和识别效率,仿真结果表明了此分类器的有效性.%Target recognition of Ballistic is a kind of multi-class target recognition. Binary tree SVM,this classification,is effective on multi-class classify. Because of simple structure and fast training it gets popularity. Moreover,it easily appears error accumulation. For resolving the flaw,a new classification is introduced in this paper. Defining adaptable degree function and adaptable intuitionistic index is to get intuitionistic fuzzy set that is used to train classification in order to get a new classification ——IFBTSVM. The new classification is applied to ballistic target recognition of HRRP which improves the correct rates and efficiency of target recognition. The results of simulation manifest the effectiveness of ITBTSVM.