摘要
提出了一种基于SURF(Speeded Up Robust Feature)的彩色图像配准算法。该算法利用图像彩色信息计算得到的颜色不变量来提取图像的特征点;提取到特征点后,结合图像的灰度信息为特征点生成特征描述子;最后采用欧氏距离进行相似性度量,提取两幅图像间匹配的特征点对。实验结果表明,该算法在保持算法的快速性和准确性的同时,获得的配准点对比SURF要多。所以该算法可以有效地避免像原始SURF算法那样因为配准点少而造成的配准失效,从而提高了算法的稳定性。
This paper proposes an automatic color image registration algorithm based on SURF（Speeded Up Robust Feature）.Firstly,a color invariant is calculated from the image’s color information to extract feature points from images by the algorithm.Secondly,after the feature points located in,the algorithm uses an image’s gray level information to build a feature descriptor.Finally,match points are extracted from a pair of images by similarity measurement criterion of Euclidean distance.The experimental results show that the algorithm can extract more match points than SURF while retaining fast computational speed and high accuracy.So it can effectively avoid registering failure due to lack of match points of the original SURF algorithm and improve stability of the algorithm.
出处
《红外技术》
CSCD
北大核心
2010年第7期415-419,共5页
Infrared Technology
基金
国家自然科学基金项目
编号:40801171
