基于SIFT特征提取的高清晰岩心图像自动配准  

SIFT-based feature extraction of high-definition core images automatic registration.

作　　者：吴晓红[1] 罗代升[1] 

机构地区：[1]四川大学电子信息学院,成都610064

出　　处：《四川大学学报（自然科学版）》2009年第1期124-128,共5页Journal of Sichuan University(Natural Science Edition)

基　　金：四川省科技攻关项目(05GG021-026-03)

摘　　要：提出了采用图像自动拼接技术获得高清晰高质量岩心扫描全景图。自动拼接的关键是特征点的提取与匹配。尺度不变特征变换算法具有伸缩、旋转、仿射不变性,并能够抗拒一定光照和视点的变化。提出把该算法用于岩心图像的特征提取,并采用欧式距离法进行特征匹配。实验结果表明,采用该算法大大提升了岩心图像拼接系统的自动化水平和准确度。This paper proposed Automatic Image Stitching technology to high-definition quality core scanning panoramic images. The key in automatic stitching is feature extraction and matching. Scale-invariant features transform algorithm with a telescopic, rotation, Affine invariant, and can resist a certain light and perspective changes. This paper proposed to apply scale-invariant features transform algorithm to image feature extraction, and used a Euclidean distance method of matching. Experimental results show that this algorithm greatly enhances the core image stitching level of automation and accuracy.

关 键 词：自动拼接 特征提取 尺度不变特征变换 欧式距离 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]