应用支持向量机分类的多角度目标识别技术  

Multi-view Target Recognition Algorithm Based on Support Vector Machine Classification

作　　者：马超杰[1] 李晓霞[1] 杨华[1] 吴丹[1] 王静雯[1] 

机构地区：[1]电子工程学院安徽省红外与低温等离子体重点实验室,安徽合肥230037

出　　处：《激光与红外》2009年第1期88-91,共4页Laser & Infrared

基　　金：安徽省红外与低温等离子体重点实验室基金(No.2007A011011F)资助

摘　　要：综合应用图像的不变矩特征和支持向量机分类方法,提出了一种对于红外图像中多角度目标的识别方法。首先通过目标分割算法求得红外图像中目标的轮廓图像,然后从轮廓图像的Hu矩、Zernike矩和Fourier-Mellin矩中选取适当阶次的矩特征组成目标在特定视角范围内的不变性特征向量;对目标的视角范围进行适当划分以解决多角度引起的目标样本多样性,并在每个划分的视角范围内分别应用支持向量机的方法进行多目标分类。测试结果表明,本文提出的方法较好地实现了红外图像中多角度目标的识别问题,是一种有效的自动目标识别算法。Synthetically utilizing image invariant moments feature and SVM （Support Vector Machine）classification method, a novel recognition algorithm was proposed to deal with multi-view target in infrared images. Firstly,the target silhouette image was extracted by image segmentation method. Then, some proper order moments which are selected from Hu moments, Zernike moments and Fourier-Mellin moments, could be used as the invariant eigen vector for target in the specific view. The angular field of view to targets was divided into several parts for solving the samples variety caused by looking from multi-view angle. In each divided field, pair-wise SVM classifier was used to realize the multitarget classification. A large number of recognition testing on multi-view targets in infrared images proves the validity and reliability of the scheme in this paper.

关 键 词：不变矩 支持向量机 自动目标识别 多视角目标识别 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]