PDF全文下载：2011040435
冯宇平， 赵文仓
（青岛科技大学 自动化与电子工程学院， 山东 青岛 266042）
摘要： 提出了结合图像局部灰度值编码匹配和相位相关进行图像模板匹配的新方法，解决了相关匹配等传统算法计算量大和已有的分块编码匹配方法因块对齐使特征比较及结果处理过程复杂的问题。匹配过程分为粗匹配和精匹配。粗匹配将模板和搜索图进行R-块划分，根据每个R-块与相邻R-块灰度值的分布关系进行编码，然后将模板在搜索图上按照步长遍历,同时对编码值进行匹配，得到初始匹配参数。接着根据初始匹配量用相位相关法进行精确匹配。实验表明,新算法抗像素灰度变化，对噪声和小角度旋转具有稳定性，其算法的时间复杂度是O（M2），比传统的归一化互相关算法快了2个数量级，与现有的分块编码方法相比速度也提高了2倍。因加入了精匹配，增强了算法的稳健性，提高了匹配定位的精度。
关键词： 模板匹配； 相似性度量； 相位相关； 时间复杂度
中图分类号： TP 391文献标志码: A
Image Matching Based on Local Gray Value Coding
FENG Yu-ping， ZHAO Wen-cang
(College of Automation and Electronic Engineering,Qingdao University of Science and Technology,Qingdao 266042,China）
Abstract: A new template matching method combining local gray value encoding matching and phase correlation is presented. The algorithm has solved the problems of large amount of calculation about the relevant matching and the complex processing caused by block alignment about existing sub-block coding matching. Matching process is divided into rough matching and fine matching. Rough matching divides the template and search image into certain size blocks called R-block，encodes the R-block according to the gray value distribution of R-block with the adjacent R-block，and matches to get the initial match parameters. Then, fine matching is finished according to the initial match. Experiments demonstrate that the new algorithm is robust to the linear transformation of pixel grey value and is stable with noise and small angle rotation. The time complexity of the proposed method is .The new algorithm is faster than traditional NCC algorithm by two orders of magnitude, and with the existing sub-block coding method has increased the speed twice. Matching precision is enhanced because of fine matching.
Key words: template matching; similarity measure; phase correlation; time complexity
收稿日期：2011-04-27
作者简介： 冯宇平（1982—），女，博士.
