一种改进的距离变换匹配算法在遥感图像目标识别中的应用  

An improved distance transform matching algorithm in remote sensing image target identification

作　　者：石磊[1] 靳敏[1] 吴长伟[1] 

机构地区：[1]黑龙江工程学院计算机科学与技术系,黑龙江哈尔滨150050

出　　处：《黑龙江工程学院学报》2010年第2期66-68,共3页Journal of Heilongjiang Institute of Technology

基　　金：黑龙江省自然科学基金资助项目(F200827)

摘　　要：将一种改进的距离变换引入到遥感图像匹配中,先对参考图像和目标图像进行直方图处理,以克服不同光照条件下带来的匹配误差,在此基础上对原图像和模板图像进行距离变换,其中引入膨胀算子,减少运算次数。最后对距离变换后的图像进行匹配操作。实验结果表明:这种方法有效地克服了几何失真、边缘变化的影响,具有抗灰度反转的能力,广泛应用于不同成像条件下获取图像间的匹配,该算法满足一定的匹配精度。The paper will introduce an improved distance transform image matching;the first goal of the reference image and histogram image processing is to overcome the matching error brought by the different lighting conditions;based on that the reference image and target images make the distance transformation.Finally,the distance transform image maks match operation.The experimental results show that this method is not only effective to overcome the geometric distortion,the edge of the impact of changes,but also has the ability to reverse anti-gray,and can be widely applied to different imaging conditions to obtain the matching between the images,and the algorithm to satisfy a certain degree of match accuracy.

关 键 词：距离变换 图像匹配 目标识别 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]