1.周口师范学院网络工程学院,河南周口 466001
2.中国科学院自动化研究所,北京 100190
[ "王 伟 男,1976年10月生于河南省周口市. 2011年获西南交通大学硕士学位,2015年获中国科学院自动化研究所博士学位.现为周口师范学院副教授. 研究方向为计算机视觉、机器学习. E-mail:wangwei@zknu.cn" ]
[ "于 磊 男,1984年6月生于河南省周口市. 2017年获河南大学硕士学位.现为周口师范学院讲师. 研究方向为计算机视觉、大数据分析." ]
收稿:2018-10-10,
修回:2019-01-10,
纸质出版:2021-08-25
移动端阅览
王伟,于磊,任国恒等.城市建筑立面三维“线-面”结构快速重建[J].电子学报,2021,49(08):1551-1560.
WANG Wei,YU Lei,REN Guo-heng,et al.Rapidly Reconstructing 3D Line-Plane Structures of Urban Building Facades[J].ACTA ELECTRONICA SINICA,2021,49(08):1551-1560.
王伟,于磊,任国恒等.城市建筑立面三维“线-面”结构快速重建[J].电子学报,2021,49(08):1551-1560. DOI: 10.12263/DZXB.20180752.
WANG Wei,YU Lei,REN Guo-heng,et al.Rapidly Reconstructing 3D Line-Plane Structures of Urban Building Facades[J].ACTA ELECTRONICA SINICA,2021,49(08):1551-1560. DOI: 10.12263/DZXB.20180752.
本文提出一种有效的基于线段匹配与结构先验的城市建筑立面三维“线-面”结构快速重建算法.本文算法利用在当前图像中检测的初始线段将图像分割为互不重叠的区域,然后在点-线约束空间内对每个区域的竖直边进行匹配并通过全局多平面拟合的方式获取建筑初始主平面;在此基础上,利用结构先验(如共线、共面等)对图像中潜在线段匹配(或空间线段)进行推断与优化,进而获取每个区域对应的空间平面.实验结果表明,本文算法仅利用两幅图像即可重建以“线-面”形式表达的城市建筑立面完整结构,整体上具有较高的效率与精度.
The paper presents an effective method to model building facades based on line segment matching and structure priors. The proposed method splits the image into several non-overlapping regions using the lines detected in the image
and then performs a rapid matching for the vertical edges of each region within a highly restrained point-line space and extracts dominant building planes using a global multi-plane fitting method. Based on the resulting planes
it further infers and optimizes potential line segment matches (or spatial lines) using structure priors (e.g.
collinearity and coplanarity) to obtain the plane associated with each region. Experimental results confirm that the proposed method can reconstruct satisfactory line-plane building facade structures from only two images
and has high accuracy and efficiency on the whole.
Fan B , Wu F , Hu Z Y . Robust line matching through line-point invariants [J]. Pattern Recognition , 2012 , 45 ( 2 ): 794 - 805 .
Kim C , Manduchi R . Planar structures from line correspondences in a manhattan world [A]. Proc of Asian Conference on Computer Vision [C]. Cham,GER : Springer , 2015 . 509 - 524 .
Al-Shahri M , Yilmaz A . Line matching in wide-baseline stereo: a top-down approach [J]. IEEE Transactions on Image Processing , 2014 , 23 ( 9 ): 4199 - 4210 .
Verhagen B , Timofte R , Gool L V . Scale-invariant line descriptors for wide baseline matching [A]. Proc of IEEE Winter Conference on Applications of Computer Vision [C]. Washington, USA : IEEE Computer Society , 2014 . 493 - 500 .
Jia Q , Gao X , Fan X , et al . Novel coplanar line-points invariants for robust line matching across views [A]. Proc of European Conference on Computer Vision [C]. Cham, GER : Springer , 2016 . 599 - 611 .
Li K , Yao J . Line segment matching and reconstruction via exploiting coplanar cues [J]. ISPRS Journal of Photogrammetry & Remote Sensing , 2017 , 125 : 33 - 49 .
Bodis-Szomoru A , Riemenschneider H , Gool L V . Fast , approximate piecewise-planar modeling based on sparse structure-from-motion and superpixels[A]. Proc of Computer Vision and Pattern Recognition [C]. Washington,USA : IEEE , 2014 . 23 - 28 .
Verleysen C , Vleeschouwer C D . Piecewise-planar 3D approximation from wide-baseline stereo [A]. Proc of Computer Vision and Pattern Recognition [C]. Washington,USA : IEEE , 2016 . 3327 - 3336 .
Tola E , Lepetit V , Fua P . Daisy: An efficient dense descriptor applied to wide-baseline stereo [J]. Pattern Analysis and Machine Intelligence , 2010 , 32 ( 5 ): 815 - 830 .
Gao W . CASIA datasets [EB/OL]. http://vision.ia.ac.cn/data/index.html http://vision.ia.ac.cn/data/index.html , 2014-12-26 .
Almazàn E J , Tal R , Qian Y , et al . MCMLSD: A dynamic programming approach to line segment detection [A]. Proc of Computer Vision and Pattern Recognition [C]. Washington, USA : IEEE , 2017 . 5854 - 5862 .
Zhao H , Shi J , Qi X , et al . Pyramid scene parsing network [A]. Proc of Computer Vision and Pattern Recognition [C]. Washington, USA : IEEE , 2016 . 6230 - 6239 .
Hartley R , Zisserman A . Multiple View Geometry in Computer Vision [M]. Cambridge,UK : Cambridge University Press , 2004 .
Delong A , Osokin A , Isack H N , et al . Fast approximate energy minimization with label costs [J]. International Journal of Computer Vision , 2012 , 96 ( 1 ): 1 - 27 .
Huang X F . Cooperative optimization for energy minimization: a case study of stereo matching [EB/OL]. http://arxiv.org/pdf/cs.CV/0701057 http://arxiv.org/pdf/cs.CV/0701057 , 2007-01-09 .
Zisserman A . VGG datasets [EB/OL]. http://www. robots.ox.ac.uk/~vgg/data/data-mview.html http://www.robots.ox.ac.uk/~vgg/data/data-mview.html , 2018-03-16 .
0
浏览量
10
下载量
3
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621