基于灰度统计的快速模板匹配算法  

Fast Template Matching Arithmetic Based on Image Gray Statistic

作　　者：陈皓[1,2,3] 马彩文[1] 陈岳承 孙小林[3] 唐自力[1,2,3] 

机构地区：[1]中国科学院西安光学精密机械研究所,西安710119 [2]中国科学院研究生院,北京100049 [3]中国华阴兵器试验中心,陕西华阴714200

出　　处：《光子学报》2009年第6期1586-1590,共5页Acta Photonica Sinica

摘　　要：针对传统的基于灰度值的匹配算法计算量大、时间复杂度高的缺陷,提出了一种基于灰度统计的快速模板匹配算法,以十字特征直线代替模板进行粗匹配,仅对搜索图中像素点数目较少的一段灰度区域进行匹配,从而大大减小了计算量,使匹配速度得到极大提高.仿真实验证明本算法匹配速度快、较传统的灰度相关算法快了二至三个数量级,具有较强鲁棒性.In order to sovle the problem that traditional matching arithmetic based on pixel gray value have defect such as huge account quantity and high time complexity,a fast template matching arithmetic based on image gray statistic was put forward. Using characteristic cross line replace template to match image, only match a gray region having less pixel amount in search image. The arithmetic reduces the account quantity,and increases the matching speed greatly. The results show that the arithmetic has high matching speed,and improves speed two or three order of magnitude in contrast to traditional matching algorithm.

关 键 词：灰度统计 模板匹配 十字特征直线 时间复杂度 搜索图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]