一种基于GLCM特征向量的目标毁伤信息提取方法  

A Target Damage Information Extraction Method Based on GLCM Feature Vector

作　　者：王威[1] 蔡超[1] 

机构地区：[1]华中科技大学自动化学院多谱信息处理技术国防科技重点实验室,武汉430074

出　　处：《计算机与数字工程》2015年第11期1946-1950,共5页Computer & Digital Engineering

摘　　要：提出了一种基于灰度共生矩阵(GLCM)特征向量的目标毁伤信息提取方法。该方法首先利用目标知识模板实现目标关键部位的快速提取,然后在进行变化区域预检测的基础上,采用GLCM提取目标被打击前后图像的纹理特征向量,使用马氏距离对纹理特征向量间的差异进行度量进一步获取变化区域,最后提取目标毁伤信息。实验结果表明,与灰度差值法和单一纹理特征变化检测方法相比,该方法能准确、有效地提取目标毁伤信息。A target damage information extraction method based on GLCM feature vector is proposed .Firstly ,the key positions of the target are extracted quickly by using the target knowledge template .After the early detection of the region with changes ,GLCM is used to extract texture features vector of the images taken before and after the attack .Then the re-gion with changes is extracted further by measuring the difference between the texture feature vectors which is based on ma-halanobis distance .Lastly ,the target damage information is extracted .Experimental results show that ,compared with the gray scale difference method and the single texture feature change detection method ,the proposed method can extract the tar-get damage information accurately and effectively .

关 键 词：变化检测 灰度共生矩阵 特征向量 毁伤信息 

分 类 号：O235[理学—运筹学与控制论]