MSEA及其在模板匹配中的应用  

MSEA AND ITS APPLICATION TO TEMPLET MATCHING

作　　者：杨静[1] 丘江[1] 刘波[1] 

机构地区：[1]中国科学院西安光学精密机械研究所,西安710068

出　　处：《光子学报》2001年第4期451-454,共4页Acta Photonica Sinica

基　　金：中国科学院西安分院院长基金资助项目

摘　　要：本文介绍了一种顺序排除算法 (SEA)及其多级算法 (MSEA) .SEA通过讨论模板的灰度信息和与局部图象的灰度信息和的关系来判断在搜索窗中的搜索点是否需要进行匹配运算 ,这样就减少了匹配估计次数 .MSEA是 SEA的多级算法 ,当搜索图背景较复杂时 ,MSEA具有更高的效率 .作为耗尽型算法 (ESA) ,SEA和 MSEA不会降低匹配的准确性 .通过仿真实验证明 ,此算法在不降低匹配准确性的前提下 ,与常用的匹配算法相比较 ,有效地减少了运算量 。In this paper,Successive Elimination Algorithm (SEA) and Multilevel Successive Elimination Algorithm (MSEA) are showed in details.In SEA,the search is performed only on the coordinate point in the search window those sum norms satisfy a certain inequality relation to the energy of the template.In this way,the operation times are much more reduced.When the background is complex,efficiency of MSEA is higher.As a Exhaustive Search Algorithm (ESA),SEA and MSEA can achieve the same estimate accuracy as original one.The efficiency and the accuracy of these algorithms are proved through the simulation result.Because of the decrease of the operation times,it is very advantageous to satisfy the requirement of real time of the system.

关 键 词：模板匹配 MSEA 顺序排除算法 图象相关匹配跟踪 灰度信息 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]