首页 | 本学科首页   官方微博 | 高级检索  
     

一种新的基于图像主轮廓的配准算法
引用本文:雷凯,刘艳滢,王延杰,邢忠宝,尹立敏.一种新的基于图像主轮廓的配准算法[J].微计算机信息,2007,23(6):267-268.
作者姓名:雷凯  刘艳滢  王延杰  邢忠宝  尹立敏
作者单位:1. 130033,吉林长春,中国科学院,长春光学精密机械与物理研究所;100039,北京,中国科学院研究生院
2. 130033,吉林长春,中国科学院,长春光学精密机械与物理研究所
摘    要:文中提出了一种新的图像配准方法,可以很好解决图像配准中的平移和旋转问题(刚性变换问题)。算法的实现是首先得到待配准两幅图像的轮廓特征点(本文使用SUSAN边缘检测算法);再剔除虚假的特征点(噪声点)后,根据像素间的连通性判别准则,得到图像的主轮廓;最后以两幅图像的主轮廓信息,配准图像。实验证明,算法实现速度快,精度较高。

关 键 词:图像配准  主轮廓  边缘检测
文章编号:1008-0570(2007)02-3-0267-02
修稿时间:2006年1月3日

A New Image Registration Algorithm Based on Main Contour
LEI KAI,LIU YANYING,WANG YANJIE,XING ZHONGBAO,YIN LIMIN.A New Image Registration Algorithm Based on Main Contour[J].Control & Automation,2007,23(6):267-268.
Authors:LEI KAI  LIU YANYING  WANG YANJIE  XING ZHONGBAO  YIN LIMIN
Affiliation:LEI KAI LIU YANYING WANG YANJIE XING ZHONGBAO YINLIMIN
Abstract:A new algorithm about image registration is presented in this paper, which has solved translations and rotations transformations (so-called rigid transformations) problems in image registration process well. This approach is as followed: first, we use SUSAN edge-detected method to obtain the feature points (interesting points) of the two images, and from which, we eliminate the bad-feature points (the noise points); second, by using connectedness-judged criteria, acquire the main outlines of the two images; third and last, we use the information extracted from the main outlines to register the two images. The experiments have test the effect of the approach.
Keywords:SUSAN
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号