«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《建筑科学与工程学报》[ISSN:1673-2049/CN:61-1442/TU]

卷:

39卷

期数:

2022年04期

页码:

81-89

栏目:

出版日期:

2022-07-12

文章信息/Info

Title:

3D Reconstruction of Buildings Based on Line Feature Optimization

文章编号:

1673-2049(2022)04-0081-09

作者:

赵红蕊¹,秦进春^1,2,谭琪凡¹

(1. 清华大学 土木工程系,北京 100084; 2. 西安测绘研究所,陕西 西安 710054)

Author(s):

ZHAO Hong-rui¹, QIN Jin-chun^1,2, TAN Qi-fan¹

(1. Department of Civil Engineering, Tsinghua University, Beijing 100084, China; 2. Xi'an Research Institute of Surveying and Mapping, Xi'an 710054, Shaanxi, China)

关键词:

建筑物;  三维重建;  直线特征;  特征检测;  特征匹配

Keywords:

building;  3D reconstruction;  straight line feature;  feature detection;  feature matching

分类号:

TU973.2

DOI:

10.19815/j.jace.2021.07138

文献标志码:

A

摘要:

为了提升建筑物三维重建精度和效率,提出基于直线特征优化的建筑物三维重建方法。首先,在分析传统直线特征检测与匹配算法的基础上,通过设定直线长度阈值,提高两视图直线特征匹配效率。随后,将密集点云引入传统直线特征匹配算法,提出点特征辅助的直线特征匹配方法,匹配直线数量明显增长。最后,结合建筑物特点,将三轴垂直关系作为约束条件引入光束法平差中,优化了建筑物模型结构,并将优化后的直线特征重采样为离散点,进一步加密建筑物点云,改善建筑物轮廓及其邻域的建模效果。结果表明:基于直线特征优化的建筑物三维重建方法能够在保证模型精度的基础上,有效改善建筑物模型结构和建模效果; 在直线特征加密试验中,优化后的直线特征能够有效加密建筑物稀疏三维点云,在密集点云中仍能有效优化加密区域的建模效果; 在模型结构优化试验中,改进后的光束法平差优化方法能够明显改善建筑物模型三轴垂直关系,有效提高了建筑物模型精度。

Abstract:

In order to improve the accuracy and efficiency of 3D reconstruction of building, a building 3D reconstruction method based on line feature optimization was proposed. Firstly, based on the analysis of the traditional line feature detection and matching algorithm, the line length threshold was set to improve the efficiency of line feature matching between two views. Then, the dense point cloud was introduced into the traditional line feature matching algorithm, and a point feature assisted line feature matching method was proposed, and the number of matching lines increased significantly. Finally, combined with the characteristics of buildings, the three-axis vertical relationship was introduced into the beam adjustment as a constraint condition to optimize the building model structure. The optimized linear features was resampled into discrete points to further encrypt the building point cloud and improve the modeling effect of building contour and its neighborhood. The results show that 3D reconstruction method based on linear feature optimization can effectively improve the structure and modeling effect of building model on the basis of ensuring the accuracy of the model. In the experiment of line feature encryption, the optimized line feature can effectively encrypt the sparse 3D point cloud of buildings, and can still effectively optimize the modeling effect of the encrypted area in the dense point cloud. In the model structure optimization experiment, the improved beam adjustment optimization method can significantly improve the three-axis vertical relationship of the building model and effectively improve the accuracy of the building model.

参考文献/References:

[1] 聂隐愚,唐 兆,常 建,等.基于单目图像的列车事故场景三维重建[J].交通运输工程学报,2017,17(1):149-158.
NIE Yin-yu,TANG Zhao,CHANG Jian,et al.3D Reconstruction of Train Accident Scene Based on Monocular Image[J].Journal of Traffic and Transportation Engineering,2017,17(1):149-158.
[2]REASON M,IAIN W,FRANCE G,et al.Structure from Motion(SfM)Photogrammetry with Drone Data:A Low Cost Method for Monitoring Greenhouse Gas Emissions from Forests in Developing Countries[J].Forests,2017,8(3):68.
[3]DANDOIS J P,BAKER M,OLANO M,et al.What Is the Point?Evaluating the Structure,Color,and Semantic Traits of Computer Vision Point Clouds of Vegetation[J].Remote Sensing,2017,9(4):355.
[4]BERGAMASCO F,TORSELLO A,SCLAVO M,et al.WASS:An Open-source Pipeline for 3D Stereo Reconstruction of Ocean Waves[J].Computers & Geosciences,2017,107:28-36.
[5]JENI L A,COHN J F,KANADE T.Dense 3D Face Alignment from 2D Video for Real-time Use[J].Image and Vision Computing,2017,58:13-24.
[6]DOU P F,WU Y H,SHAH S K,et al.Monocular 3D Facial Shape Reconstruction from a Single 2D Image with Coupled-dictionary Learning and Sparse Coding[J].Pattern Recognition,2018,81:515-527.
[7]XIE L,ZHANG X H,XU Y H,et al.Skeleton Fusion:Reconstruction and Tracking of Human Body in Real-time[J].Optics and Lasers in Engineering,2018,110:80-88.
[8]ZHU H,LIU Y B,FAN J T,et al.Video-based Outdoor Human Reconstruction[J].IEEE Transactions on Circuits and Systems for Video Technology,2017,27(4):760-770.
[9]SOURIMANT G,MORIN L,BOUATOUCH K.GPS,GIS and Video Registration for Building Reconstruction[C]//IEEE.Proceedings of 2007 IEEE International Conference on Image Processing.San Antonio:IEEE,2007:401-404.
[10]ZHAO J H,LI D R,WANG Y M.Ancient Architecture Point Cloud Data Segmentation Based on Modified Fuzzy C-means Clustering Algorithm[C]//HUANG J,QI P,WAN Y,et al.Proceedings of International Conference on Earth Observation Data Processing and Analysis.Wuhan:SPIE,2008:469-478.
[11]USLU B,GOLPARVAR-FARD M,DE LA GARAZA J M.Image-based 3D Reconstruction and Recognition for Enhanced Highway Condition Assessment[C]//ZHU Y,RAYMOND R.Proceedings of International Workshop on Computing in Civil Engineering 2011.Miami:ASCE,2011:67-76.
[12]MORENO F A,GONZALEZ-JIMENEZ J,BLANCO J L,et al.An Instrumented Vehicle for Efficient and Accurate 3D Mapping of Roads[J].Computer-aided Civil and Infrastructure Engineering,2013,28(6):403-419.
[13]LATTANZI D,MILLER G R.3D Scene Reconstruction for Robotic Bridge Inspection[J].Journal of Infrastructure Systems,2015,21(2):04014041.
[14]张 恒,余 涛,柳 鹏,等.高分辨率 SAR 影像形态学层级分析的建筑物检测[J].中国图象图形学报,2015,20(11):1517-1525.
ZHANG Heng,YU Tao,LIU Peng,et al.Building Detection of High-resolution SAR Image Based on Morphological Hierarchy Analysis[J].Journal of Image and Graphics,2015,20(11):1517-1525.
[15]张 浩,赵云胜,陈冠宇,等.基于支持向量机的遥感图像建筑物识别与分类方法研究[J].地质科技情报,2016,35(6):194-199.
ZHANG Hao,ZHAO Yun-sheng,CHEN Guan-yu,et al.Remote Sensing Image Building Recognition and Classification Based on the Support Vector Machine[J].Geological Science and Technology Information,2016,35(6):194-199.
[16]王燕妮,王慧琴,王建平.局部特征结合形状轮廓匹配的建筑物识别算法[J].西安建筑科技大学学报(自然科学版),2017,49(5):752-756.
WANG Yan-ni,WANG Hui-qin,WANG Jian-ping.A Building Recognition Algorithm Based on Local Feature and Shape Contour Matching[J].Journal of Xi'an University of Architecture and Technology(Natural Science Edition),2017,49(5):752-756.
[17]王思洁,方莉娜,陈崇成,等.基于结构化场景的单幅图像建筑物三维重建[J].地球信息科学学报,2016,18(8):1022-1029.
WANG Si-jie,FANG Li-na,CHEN Chong-cheng,et al.A Single View Type 3D Reconstructive Method for Architecture Based on Structured Scene[J].Journal of Geo-Information Science,2016,18(8):1022-1029.
[18]杨 玲,阮心玲.模型导向的建筑物半自动三维重建[J].遥感信息,2017,32(3):115-122.
YANG Ling,RUAN Xin-ling.Semi-automatic Model-driven 3D Building Reconstruction[J].Remote Sensing Information,2017,32(3):115-122.
[19]VON GIOI R G,JAKUBOWICZ J,MOREL J M,et al.LSD:A Line Segment Detector[J].Image Processing on Line,2012,2:35-55.
[20]QI J,GAO X K,XIN F,et al.Novel Coplanar Line-points Invariants for Robust Line Matching Across Views[C]//LEIBE B.Proceedings of 14th European Conference on Computer Vision.Amsterdam:Springer,2016:599-611.
[21]MOURAGNON E,LHUILLIER M,DHOME M,et al.Generic and Real-time Structure from Motion Using Local Bundle Adjustment[J].Image and Vision Computing,2009,27(8):1178-1193.

相似文献/References:

备注/Memo

备注/Memo:

收稿日期:2021-07-21
基金项目:国家重点研发计划项目(2018YFD1100905)
作者简介:赵红蕊(1969-),女,河北唐山人,教授,博士研究生导师,理学博士,E-mail:zhr@tsinghua.edu.cn。

常用功能

更新日期/Last Update: 2022-07-10