具有仿射不变性的彩色图像配准方法研究  

A Method of Affine Invariant for Color Image Registration

作　　者：黄友文[1] 黄雅兰[1] 袁春经 

机构地区：[1]江西理工大学信息工程学院,江西赣州341000

出　　处：《小型微型计算机系统》2015年第2期338-342,共5页Journal of Chinese Computer Systems

基　　金：江西省教育厅科技基金项目(GJJ10156)资助;江西理工大学校级课题项目(NSFJ2014-K18)资助

摘　　要：经典SIFT算法在特征提取及特征匹配阶段存在耗时过长的问题,而SURF算法对仿射角度大的图像配准时错配率较高;且已有算法对彩色图像的处理均把彩色信息转为灰度信息,从而丢失了表征图像的颜色信息.提出一种具有仿射不变性的彩色图像配准算法:C-ASURF算法,通过模拟仿射变换得到一系列模拟图像,对模拟得到的图像计算颜色不变量,再对其进行特征点的检测及特征点的匹配,因此,它能够很好的解决图像配准时因图像仿射角度过大而造成的配准失效问题,且能够保留彩色图像的彩色信息.实验结果表明,C-ASURF算法的运行速度比具有仿射不变性的SIFT算法要快很多,且图像配准的准确率也很高,能够满足图像配准的高效性及实时性要求,具有很好的稳定性.The classical SIFT algorithm takes too long time in the feature extraction and feature matching phase, and SURF algorithm fails to image registration for big affine angles of images, and the mismatch rate is very high. Most of proposed algorithms change col- or image into gray image, thus which lost the color characterization for the color image. A method of the affine invariance for color im- age registration （ C-ASURF ） is proposed in this paper i which obtains a series of analog images by simulate the different affine angles, then calculates the color invariants and extracts features then match the features. Thus the C-ASURF algorithm can solve the mismatch problem for images with large affine angle, and can retain the color information for the color image. The experimental results show that C-ASURF algorithm runs faster than ASIFT algorithm （ affine invariant SIFT） ,and the match accuracy rate is quite high,and can meet the real-time requirements and the high efficiency requirements of image registration, and has good stability.

关 键 词：仿射不变性 特征提取 颜色不变量 图像配准 实时性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]