基于图像增强技术的SURF特征匹配算法研究  

Research on SURF Feature Matching Algorithm Based on Image Enhancement Technology

作　　者：张明浩 杨耀权[1] 靳渤文 ZHANG Ming-hao;YANG Yao-quan;JIN Bo-wen(School of Control and Computer Engineering,North China Electric Power University,Baoding 071003,China)

机构地区：[1]华北电力大学控制与计算机学院

出　　处：《自动化与仪表》2019年第9期98-102,共5页Automation & Instrumentation

摘　　要：图像匹配是数字图像研究中的一个重点和难点,SURF算法是图像特征点提取与匹配常用算法之一,是对SIFT算法的改进,提升了算法的执行效率,为算法在实时计算机视觉系统中应用提供了可能。该文采用SURF算法提取图像特征,增加RANSAC算法来对匹配点进行筛选。但对于光照影响导致曝光不足或曝光过度图像的匹配效果欠佳,据此提出将图像增强技术与SURF算法相结合的思想,以得到更好的特征点质量以及匹配效率,并针对不同特征的图像采用适当的图像增强方法,利用实验进行了对比和验证。Image matching is an important and difficult point in digital image research. SURF algorithm is one of the commonly used algorithms for image feature point extraction and matching. It is an improvement of SIFT algorithm,which improves the execution efficiency of the algorithm. The application in the system provides the possibility. In this paper,the SURF algorithm is used to extract image features,and the RANSAC algorithm is added to filter the matching points. However,the effect of illumination on the underexposed or overexposed image is not good. Based on this,the idea of combining image enhancement technology with SURF algorithm is proposed to obtain better feature point quality and matching efficiency,and to target different features. The images were compared and verified using experiments using appropriate image enhancement methods.

关 键 词：图像匹配 SURF算法 图像增强 RANSAC算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]