红外小目标图像的背景杂波量化方法  

Background Clutter Quantification Method for Infrared Image of Small Targets

作　　者：黄康[1] 毛峡[1] 梁晓庚[1,2] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100191 [2]中国空空导弹研究院,河南洛阳471009

出　　处：《光学学报》2011年第3期57-62,共6页Acta Optica Sinica

基　　金：航空科学基金(20070112001);航天支撑技术基金资助课题

摘　　要：分析了背景杂波对红外小目标检测的影响,研究了一种针对红外小目标图像的背景杂波定量描述方法。提出了背景杂波定量描述方法需要满足的三个条件:与主观判断的结果一致、对于不同的红外图像具有适应性和能够辅助改进目标检测算法。综合考虑了目标与背景的特征,融合了4种特征,提出了一种新的红外背景杂波量化方法。首先采用支持向量机对背景杂波强度进行分类,然后对每类图像分别采用层次分析法计算各特征的权重,从而实现杂波强度的量化。设计了与传统方法的对比实验,实验结果表明该方法是一种有效的背景杂波强度量化方法。By analyzing the influence of background clutter on infrared small targets detection,we have researched a background clutter description method for infrared images of small targets.Three conditions of a background clutter description method are proposed： consistent with subjective judgment;suituble for different infrared images;assistant with improvement of target detection algorithms.Four features are synthesized by considering the relation between targets and backgrounds,and a novel method is proposed for describing the background clutter quantitatively in infrared images of small targets.Firstly,supporting vector machine is emploged to classify the backgrounds.Secondly,the weight of each feature is calculated by analytic hierarchic process for each class of background images.Then the quantification problem is solved.Compared with the conventional methods,the validity of the proposed method has been demonstrated by the experimental results.

关 键 词：图像处理 红外图像 小目标 背景杂波 定量描述 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]