Matching extension for structured scenes based on homography constraint

ZHAI You; XIONG Wei; ZENG Luan

doi:10.3788/OPE.20152313.0466

您当前的位置：

首页 >

文章列表页 >

Matching extension for structured scenes based on homography constraint

更新时间：2020-08-13

- Matching extension for structured scenes based on homography constraint
- Optics and Precision Engineering Vol. 23, Issue 10z, Pages: 466-471(2015)
- 作者机构：
  
  中国人民解放军装备学院北京,101416
- 作者简介：
- 基金信息：
- DOI：10.3788/OPE.20152313.0466
  CLC： TP751
- Received：10 April 2015，
  
  Revised：05 May 2015，
  
  Published：14 November 2015
- 稿件说明：
移动端阅览
翟优, 熊伟, 曾峦. 利用单应约束实现结构化场景的匹配扩展[J]. 光学精密工程, 2015,23(10z): 466-471

ZHAI You, XIONG Wei, ZENG Luan. Matching extension for structured scenes based on homography constraint[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 466-471
翟优, 熊伟, 曾峦. 利用单应约束实现结构化场景的匹配扩展[J]. 光学精密工程, 2015,23(10z): 466-471 DOI： 10.3788/OPE.20152313.0466.

ZHAI You, XIONG Wei, ZENG Luan. Matching extension for structured scenes based on homography constraint[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 466-471 DOI： 10.3788/OPE.20152313.0466.

摘要

针对结构化场景图像纹理单一

宽基线情况下基于局部特征的匹配方法能够获取的匹配较少的问题

提出了一种基于初始匹配的匹配扩展方法。首先

依据单应约束将初始匹配划分为多个子类

每个子类对应于不同的局部平面

并采用RANSAC算法估计每个子类的单应矩阵

剔除不符合平面约束的外点。然后

依据平面单应搜索未匹配特征点的候选匹配以及改进8扇区SURF描述符的相似性剔除错误匹配。最后

在扩展后的匹配集合中更新单应估计

并剔除外点

实现匹配扩展。实验结果表明:当图像存在一定的尺度和视点变化以及明显的光照变化时

该方法能够有效地增加匹配的数目

匹配增长率均在100%以上

可为精确的极线几何估计和三维重建奠定良好的基础。

Abstract

Structured scenes have uniform texture and the matching method based on local features can get few matches in a wide baseline case. Therefore

a matching extension method based on homography constraint was presented. Firstly

initial matches were divided into several groups and each group was corresponded to different planes based on the homography constraints. The homography matrix for each group was estimated using RANSAC based strategy and the outliers out of the plane constraint were removed. Then

the nearest candidate correspondences for unmatched features such as corners and blobs were searched by homography transformation and the SURF descriptor was used to further reject outliers. Furthermore

the homography constraints were updated using extended matches and outliers are removed. Experiment results show that the proposed method can adapt to a certain degree of scale and viewpoint changes and obviously illumination change. The initial matches are greatly enriched

and matching growth rates all achieve to 100%

which lays a good foundation for accurate epipolar geometry estimation and three dimensional reconstruction.

关键词

Keywords

references

MATAS J, CHUM O, URBAN M, et al.. Robust Wide Baseline Stereo from Maximally Stable Extremal Regions[J]. Image and Vision Computing, 2004, 22(10):761-767.

MIKOLAJCZYK K, SCHMID C. An affine invariant interest point detector[C]. Proceedings of the 7th European Conference on Computer Vision, Denmark, Springer, 2002:128-142.

MIKOLAJCZYK K, SCHMID C. Scale & affine invariant interest point detectors[J]. International Journal on Computer Vision, 2004, 60(1):63-86.

LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2):9l-110.

HERBERT B, ANDREAS E, TINNE T, et al.. Speeded up robust features[J]. Computer Vision and Image Understanding, 2008, 110(3):346-359.

WANG Z H, ZHI S S, LIU H M. Mshs:The mean-standard deviation curve matching algorithm in hsv space[C]. Proceedings of International Conference on Machine Learning and Cybernetics, Xian, IEEE, 2012, 3:1064-1069.

ZHANG L, KOCH R. Line matching using appearance similarities and geometric constraints[C]. Proceedings of DAGM/OAGM Symposium, Austria, Springer, 2012:236-245.

VERHAGEN B, TIMOFTE R, GOOL L V. Scale-Invariant Line Descriptors for Wide Baseline Matching[C]. Proceedings of IEEE Winter Conference on Applications of Computer Vision, Colorado, IEEE, 2014:493-500.

LU W, ULRICH N, SUYA Y. Wide-baseline image matching using Line Signatures[C]. Proceedings of IEEE 12th International Conference on Computer Vision, Kyoto, IEEE, 2009:1311-1318.

曾峦, 王元钦, 谭久彬. 改进的SIFT特征提取和匹配算法[J]. 光学精密工程, 2011, 19(6):1391-1397. ZENG L, WANG Y Q, TANG J B. Improved algorithm for SIFT feature extraction and matching[J]. Opt. Precision Eng., 2011, 19(6):1391-1397.

KANNALA J, BRANDT S. Quasi-dense wide baseline matching using match propagation[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis:IEEE,2007:1-8.

翟优, 曾峦, 熊伟. 不同局部邻域划分加速鲁棒特征描述符的性能分析[J]. 光学精密工程, 2013, 21(9):2395-2404. ZENG Y, ZENG L, XIONG W. Performance analysis of SURF descriptor with different local region partition[J]. Opt. Precision Eng., 2013, 21(9):2395-2404.

翟优, 熊伟, 曾峦. 利用直线交点特征实现宽基线匹配.(外审) ZENG Y, XIONG W, ZENG L. Wide-baseline Matching Based on Line Intersection Features.(Under review)

Views

291

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Performance analysis of SURF descriptor with different local region partition

Improved algorithm for SIFT feature extraction and matching

Research of 3D reconstruction and position estimation of 3D pieces based on binocular structured light peripheral scanning

Design of partial overlap point cloud registration network driven by overlap score and matching matrix

A method for underwater vehicle attitude measurement based on Snell’s window and its polarization mode

Related Author

ZHAI You

ZENG Luan

XIONG Wei

ZENG Luan

WANG Yuan-qin

TAN Jiu-bin

CHENG Yao

WU Zhetao

Related Institution

Key Laboratory, Academy of Equipment

哈尔滨工业大学超精密光电仪器工程研究所

装备指挥技术学院

College of Mechanical Engineering， Chongqing University of Technology

School of Information Engineering， Jiangxi University of Science and Technology

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰