基于Harris角点匹配的工具识别方法  

Tool recognition method based on Harris corner matching

作　　者：赵柏山[1] 肖乃瑶 张学松[1] 

机构地区：[1]沈阳工业大学信息科学与工程学院,辽宁沈阳110000

出　　处：《微型机与应用》2017年第6期43-45,50,共4页Microcomputer & Its Applications

基　　金：辽宁省教育厅一般研究项目(L2014053)

摘　　要：图像匹配对于工具识别过程具有重要意义,传统的图像匹配算法匹配过程复杂导致匹配速度慢,在实际应用中有局限。为了提高工具识别的效率,通过分析工具图像特征,设计出一种基于Harris角点特征的图像匹配方法进行工具识别。首先用Harris算法提取工具模板图像与搜索图像角点特征,然后使用归一化互相关匹配算法计算工具模板图像与搜索图像角点特征的相关值,进而确定匹配点对,最后采用RANSAC去除错误匹配点对。通过实际数据验证,基于Harris角点特征的图像匹配方法与传统方法相比,不仅识别速度快,鲁棒性好,而且在实际工程应用中更具适用性。Image matching has great significance for tool identification process. Traditional matching algorithm of image matching algorithm is complex,which leads to slow matching speed and has limitation in practical application. In order to improve the efficiency of the tool identification,an image matching method based on Harris corner feature was designed to recognition tool by analyzing the image features of the tool.Firstly,it extracts tool template image and searches image corner point feature with Harris algorithm. And then it uses the normalized cross-correlation matching algorithm to calculate the relative value of tool template image and search image corner point features,which determine the matching point pairs. Finally,it uses RANSAC to eliminate error matching point pairs. Through the actual data validation,compared with the traditional method,the image matching method based on Harris corner feature is not only faster and more robust,but also has more applicability in practical engineering application.

关 键 词：图像匹配 工具图像特征 工具识别 Harris角点特征 归一化互相关 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]