标准互相关的灰度无关特性及其应用  

Characteristics of Independence on Image Gray Level in Pattern Matching Algorithm and its Application

作　　者：孙明磊[1] 宗光华[1] 毕树生[1] 于靖军[1] 

机构地区：[1]北京航空航天大学机械工程及自动化学院,北京100083

出　　处：《光子学报》2009年第2期435-440,共6页Acta Photonica Sinica

基　　金：国家863高技术研究与发展计划(2004AA404260);中国博士后科学基金(20070420287)资助

摘　　要：对用于图像特征定位的标准互相关算子灰度无关特性展开了研究.由于在显微视觉中的大多数图像灰阶较低,接近于三值甚至二值图像;可籍此灰度无关特性对模板匹配的算法进行优化.实验结果表明,标准互相关的灰度无关特性可以在实际中指导设计模板图像,降低相关运算计算速度,以及指导改善照明条件.A characteristic that Normalized Cross Correlation Operating （NCCO） results independent of image gray levels in pattern matching for image feature positioning was presented. In the field of microscopic vision research,the majority of images can be approximated to an image with three gray levels, or even binary image. So with this attractive feature, a method of improved pattern matching algorithms was presented. Experimental results show that the characteristic of gray level independence can give guidance to designing template images, reducing time consumption of calculations, and improving illuminating conditions.

关 键 词：图像 模式匹配 相关 灰度 无关 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置]