运用旋转无关特征线实现景象匹配

张叶; 曲宏松; 王延杰

doi:null

您当前的位置：

首页 >

文章列表页 >

运用旋转无关特征线实现景象匹配

信息科学 | 更新时间：2020-08-12

- 运用旋转无关特征线实现景象匹配
- Implementation of scene matching based on rotation invariant keylines
- 光学精密工程 2009年17卷第7期页码：1759-1765
- 作者机构：
  
  中国科学院长春光学精密机械与物理研究所,吉林长春,中国,130033
- 作者简介：
- 基金信息：
  
  某军口863项目(No.2006AA703405F)()
- DOI：
  中图分类号： TP391.4
- 收稿日期：2008-07-28，
  
  修回日期：2008-10-08，
  
  网络出版日期：2009-07-25，
  
  纸质出版日期：2009-07-25
- 稿件说明：
移动端阅览
张叶, 曲宏松, 王延杰. 运用旋转无关特征线实现景象匹配[J]. 光学精密工程, 2009,17(7): 1759-1765

ZHANG Ye, QU Hong-song, WANG Yan-jie. Implementation of scene matching based on rotation invariant keylines[J]. Editorial Office of Optics and Precision Engineering, 2009,17(7): 1759-1765
张叶, 曲宏松, 王延杰. 运用旋转无关特征线实现景象匹配[J]. 光学精密工程, 2009,17(7): 1759-1765 DOI：

ZHANG Ye, QU Hong-song, WANG Yan-jie. Implementation of scene matching based on rotation invariant keylines[J]. Editorial Office of Optics and Precision Engineering, 2009,17(7): 1759-1765 DOI：

摘要

为了寻求一种高效可靠的景象匹配算法

使其在复杂背景下具有尺度无关

抗旋转、光照变化和图像轻微畸变的特性

对基于特征线的景象匹配算法进行了研究。受人眼视觉系统(HVS)的启发

利用类似人眼视觉皮层滤波器组的处理方式

提出一种新的基于特征线的景象匹配方法。该方法利用向量描述特征线来完成图像间的匹配

相比特征点的匹配算法具有特征数量少

信息丰富的特点。介绍了利用提出的方法进行特征线的提取、表征与匹配的过程

总结了该匹配方法的效果和特点。实验表明

这种基于特征线的景象匹配算法可以完成复杂背景下图像间的匹配

克服了旋转和尺度缩放

并具有很好的鲁棒性

匹配精度优于1 pixel

满足了自动景象匹配的稳定可靠、精度高、抗干扰能力强等要求。

Abstract

In order to provide a robust scene matching algorithm with the scale and rotation invariance and less influence by different illumination under complex conditions

this paper researches scene matching techniques based on key lines. Inspired by the Human Visual System (HVS)

a new scene matching method based on key lines is put forward by using a set of filters that are the same processing models as the human cortex filter.The mathod uses the vectors to describe key lines to implement the metching between the two images and is characterized by less numbers of key line and great information as compared with that of a point matching method. It introduces the extraction

expression and matching lines and gives the simulation results and matching properties. Experiment results show that the proposed method works well for the condition of rotating and scaling

and can offer the precision of scene matching of 1 pixel

which meets the requirements of high stability and good robust for complex background scene matching.

关键词

Keywords

references

FORSYTH D A,PONCE J.计算视觉一种现代方法[M]. 北京:电子工业出版社,2004:11,89. FORSYTH D A,PONCE J.Computer Vision: A Modern Approach[M].Beijing:Publishing House of Electronics Industry,2004:11,89.(in Chinese)[2] BAILEY J G, MESSNER R A. Log-polar mapping as a preprocessing stage for an image tracking system [J]. Intelligent Robots and Computer Vision, 1988(7):223-227.[3] LIM F L,WEST G A W,VENKATESH S. Use of log polar space for foveation and feature recognition[J].IEEE, Proc. Vis. Image Signal Process,1997,144(6):323-331..[4] CHEN Q S, DEFRISE M, DECONINCK F.Symmetric phase-only matched filtering of Fourier-Mellin transforms for image registration and recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1994,16(12):1156-1168.[5] HAROLD S S,TAO B,MCGUIRE M. Analysis of image registration noise due to rotationally dependent aliasing[J].Visual Communication and Image Representation, 2003,14:14-135.[6] DAVID G L. Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision, 2004,60(2):91-110.[7] DAVID G. Lowe, local feature view clustering for 3D object recognition .Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2001,1:682-688.[8] DAVID G L. Object recognition from local scale-invariant features .International Conference on Computer Vision, 1999,2:1150-1157.[9] ZHANG Y, WANG Y J. High accuracy real-time automatic thresholding for centroid tracker[J]. SPIE,2006,6027:1-7.[10] 高莹莹,杨建峰,马晓龙,等. 基于Fourier-Mellin算法的干涉图像配准[J]. 光学精密工程,2007,15(9):1415-1420. GAO Y Y, YANG J F, MA X L,et al.. Interference image registration based on Fourier-Mellin algorithm[J]. Opt. Precision Eng.,2007,15(9):1415-1420.

浏览量

641

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

结合MSA与方向编码的快速景象匹配