基于改进模板匹配算法的伤票识别与定位  

Injured ticket detection and location based on improved template matching method

作　　者：蒋昭颖 郝福珍[1] 

机构地区：[1]华北计算技术研究所后勤信息化事业部,北京100083

出　　处：《电子设计工程》2018年第3期175-179,共5页Electronic Design Engineering

摘　　要：针对纸质伤票信息的自动化采集,提出一种改进模板匹配算法的识别与定位方法。用于移动终端的CCD相机采集具有几何特征的伤票图像,采用改进相关系数匹配的算法进行快速识别伤票并且粗定位伤票位置,根据粗定位的位置创建ROI区域,对该区域图像进行自适应阈值消除光照阴影,利用数学形态学闭运算提取边缘几何特征精确定位伤票位置,计算定位伤票与模板伤票的匹配度。实验结果表明,改进的相关系数匹配算法能够提高匹配速度,并具备较强的自适应性,数学形态学算法能够根据伤票边缘几何特征精确定位伤票位置,匹配度达到90%以上。Aiming at the problem of identifying and locating the injured ticket which exists in the injured ticket information detection and extraction,this paper proposes an improved template matching algorithm to identify and locate the injured ticket.Geometric features of injured ticket images are collected by using CCD camera,then it uses the improved correlation coefficient matching algorithm to identify ticket quickly,and locate the area approximately.According to the approximate area to create ROI area,and uses the adaptive threshold to extract the geometric features of edge,and then uses the morphological method to locate injured ticket accurately.At last,computes the matching degree of locating ticket and templet ticket.The results of the experiment show that the improved correlation coefficient matching algorithm can improve the matching efficiency,and have strong stability.Morphological method can accurately locate the injured ticket according to the edge feature.The degree of matching is above 90%.

关 键 词：图像识别 相关系数匹配 数学形态学 模板匹配 

分 类 号：TN919.8[电子电信—通信与信息系统]