基于改进的SURF的图像匹配查重算法  

Duplicate Checking Algorithm of Image Matching Based on Improved SURF

作　　者：丁一[1] DING Yi

机构地区：[1]无锡城市职业技术学院,江苏无锡214000

出　　处：《科技创新与应用》2020年第32期25-27,31,共4页Technology Innovation and Application

基　　金：2019年江苏省职业教育教学改革研究课题(编号:ZYB345)。

摘　　要：随着新型冠状病毒的蔓延,各大高校都普遍尝试和采用了线上教学的方式进行授课和评价。目前各高校普遍实行的过程化考核作为课程分数的评价标准之一。传统的查重工具着重于文字的重复率,忽视了图片这一关键的信息载体,因此急需以图像识别匹配技术作为基础的图像查重算法。文章将SURF算法应用于学生作业及实验报告等文本评价载体中的图片相似度匹配上,结合平时的实践经验,用RANSAC算法去掉错误的匹配结果,匹配算法对于SURF特征点进行优化,从而实现了对SIFT算法匹配速度以及精确度的改善,最终实现了完善的实验报告图像匹配算法,并且对实验中出现的问题进行讨论和总结,对系统实施的改进和未来的拓展性也进行了充分的论述。With the popularity of novel coronavirus,colleges and universities have generally tried and adopted online teaching and evaluation.At present,the process assessment,which is widely implemented in colleges and universities,is one of the evaluation criteria of curriculum scores.The traditional duplicate checking tools focus on the repetition rate of the text,ignoring the picture as a key information carrier,so there is an urgent need for an image repetition checking algorithm based on image recognition and matching technology.In this paper,the SURF algorithm is applied to the image similarity matching in the text evaluation carriers such as students'homework and experimental reports,combined with the usual practical experience,the wrong matching results are removed by the RANSAC algorithm,and the matching algorithm is optimized for the SURF feature points,thus the matching speed and accuracy of the SIFT algorithm are improved,and finally a perfect experimental report image matching algorithm is realized.And the problems in the experiment are discussed and summarized,and the improvement of the implementation of the system and the expansion in the future are also fully discussed.

关 键 词：SURF算法 图像查重 图片匹配度 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]