改进的多尺度形状模板匹配算法  

Improved multi-scale shape template matching algorithm

作　　者：王凯[1] 余振军 何显辉 夹尚丰 孙洋 孙林[1] WANG Kai;YU Zhenjun;HE Xianhui;JIA Shangfeng;SUN Yang;SUN Lin(College of Geodesy and Geomatics,Shandong University of Science and Technology,Qingdao Shandong 266000,China;Star-rising Technologies Co.,Ltd.,Qingdao Shandong 266000,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛266000 [2]青岛星科瑞升信息科技有限公司,山东青岛266000

出　　处：《激光杂志》2022年第4期82-87,共6页Laser Journal

基　　金：国家自然科学基金(No.41771408)。

摘　　要：针对形状模板匹配算法在目标发生尺度变化时匹配效率低的问题,提出一种改进的多尺度形状模板匹配算法。离线过程中,针对参考点和多尺度模板与目标的匹配不适应情况,提出一种新的参考点选取方法;在线过程中,采用一种优先匹配策略,确定该尺度层存在最佳模板的期望值,并依此判断该尺度层模板参与匹配的必要性;基于选取的最优参考点和优先匹配策略实现目标的高效匹配。实验结果显示,与经典算法相比,所提算法的平均匹配时间分别为NCC算法、SSDA算法的0.81%、2.25%;且当目标发生平移、旋转和尺度等复杂变化时,依然可以实现目标的精确匹配。Aiming at low matching efficiency of shape template matching algorithm when target scale changes, an improved multi-scale shape template matching algorithm is proposed. In off-line process, a new reference point selection method is proposed for the mismatch between reference point and multi-scale template and target;In online process, a priority matching strategy is adopted to determine expected value of the best template in scale layer, and judge the necessity of the scale layer template participating in the matching;Based on the selected optimal reference point and priority matching strategy, the efficient target matching is realized. The experimental results show that the average matching time of the proposed algorithm are 0.81% and 2.25% of that of NCC algorithm and SSDA algorithm respectively;And when target has complex changes such as translation, rotation and scale, the accurate matching of target can still be achieved.

关 键 词：模板匹配 边缘点 多尺度 参考点 相似度 

分 类 号：TN209[电子电信—物理电子学]