基于暗区连通域的定位孔识别方法  

Positioning holes recognition based on dark connected areas

作　　者：肖刚[1] 夏通[1] 张元鸣[1] 陈久军[1] 徐俊[2] 

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310023 [2]浙江工业大学艺术学院,杭州310023

出　　处：《计算机工程与应用》2011年第34期215-218,共4页Computer Engineering and Applications

基　　金：浙江省自然科学基金(No.Y1090903)

摘　　要：为了在普通扫描仪上实现自动化线拍系统,提出了一种基于暗区连通域的定位孔识别方法。介绍了暗区连通域的可分割性原理,该原理是定位孔连通域被准确分割的基础,也是定位孔能被正确识别的前提。该方法先依据该原理计算出连通域的分割阀值,然后对定位孔的形状及其相互位置关系进行了特征建模,利用模板匹配方法实现定位孔识别。对总共83张动画稿扫描图像进行定位孔识别测试,实验结果表明,定位孔识别正确率98.80%,对于1740万像素的动画稿图片,识别速度也能达到15.27ms/page。To implement an automatic line-test system on general scanner,this paper proposes a method of positioning holes recognition based on dark connected areas.This paper introduces a segmentation principle for dark connected areas,which is the basis of connected areas to be correctly segmented and is also the basis of positioning holes to be correctly recognized. The segmentation threshold is calculated.Then a model is built for shape and relative position of positioning holes.At last a template matching method is used to recognize positioning holes.The experimental results show that the accuracy of recognition based on this method is 98.80%,and speed reaches 15.27 ms/page on 17.4 million pixels pictures.

关 键 词：暗区连通域 模板匹配 定位孔识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]