J4 ›› 2011, Vol. 29 ›› Issue (02): 164-.

• 论文 • 上一篇    下一篇

基于特征点的匹配拼接算法研究与实现

贺 琨|张凤春|孟祥磊|宋占伟   

  1. 吉林大学 电子科学与工程学院| |长春 130012
  • 出版日期:2011-03-25 发布日期:2011-04-25
  • 通讯作者: 贺琨(1986— ),男,湖南邵阳人,吉林大学硕士研究生,主要从事图像处理研究,(Tel)86-13756523923(E-mail)hekun09@mailjlueducn E-mail:hekun09@mail.jlu.edu.cn
  • 作者简介:贺琨(1986— )|男|湖南邵阳人|吉林大学硕士研究生|主要从事图像处理研究|(Tel)86-13756523923(E-mail)hekun09@mail.jlu.edu.cn;宋占伟(1962— )|男|长春人|吉林大学教授|硕士研究生导师|主要从事数字信号处理、图像处理、智能交通、嵌入式系统研究,(Tel)86-431-85095828;85894687(E-mail)songzw@mail.jlu.edu.cn。

Research and Implementation on Image Mosaic Based on Feature Points Matching

HE |Kun|ZHANG Feng-chun|MENG Xiang-|SONG Zhan-wei   

  1. College of Electronic Science and Engineering|Jilin University|Changchun 130012, China
  • Online:2011-03-25 Published:2011-04-25

摘要:

针对图像间因具有旋转及光线强度差异等现象而导致的拼接效果不佳及拼接速度慢的问题,提出一种基于特征点的配准算法。该算法首先利用相位相关法确定图像重叠区域,然后采用改进Harris角点检测算法检测角点,再根据相似测度NCC(Normalized Cross Correlation)方法提取出匹配特征点对,最后用渐进渐出的方法实现拼接图像的融合。实验证明,该算法比传统算法在角点数目上减少了四分之一,有效地去除了拼接产生的鬼影现象。能有效地提高图像拼接的速度和精度。

关键词: 图像拼接, 特征点, 相位相关, 图像融合

Abstract:

In order to solve the problem of poor effect in image stitching caused by the rotation, and the differences in light intensity, an image stitching algorithm based on the feature points is proposed in this paper. The algorithm first determines the overlapped area in images using the method of phase correlation, then extracts the corners with improved Harris detecting algorithm,and the feature point matching pairs acquired by NCC(normalized cross correlation) method, finally we are able to complement the fusion of stitching images by using the step-in-step-out method. The experiment results show that the algorithm improves the speed and accuracy of image stitching effectively.

Key words: image mosaic, feature points, phase correlation, image fusion

中图分类号: 

  • TP311.5