基于非平衡最优传输理论的拉索式幕墙面板尺寸检测方法

doi:10.11996/JG.j.2095-302X.2025051134

图学学报 ›› 2025, Vol. 46 ›› Issue (5): 1134-1143.DOI: 10.11996/JG.j.2095-302X.2025051134

基于非平衡最优传输理论的拉索式幕墙面板尺寸检测方法

谭立云¹^,²(), 刘界鹏¹^,²(), 李汉涛³, 曾焱¹^,², 廖岳¹^,², 吴小峰³, 崔娜¹^,²

¹ 重庆大学土木工程学院，重庆 400045
² 重庆大学山地城镇建设与新技术教育部重点实验室，重庆 400045
³ 广州葛洲坝建设工程有限公司，广东广州 511466

收稿日期:2024-11-11 接受日期:2025-04-25 出版日期:2025-10-30 发布日期:2025-09-10
通讯作者:刘界鹏(1978-)，男，教授，博士。主要研究方向为智能建造。E-mail：liujp@cqu.edu.cn
第一作者:谭立云(2001-)，女，硕士研究生。主要研究方向为智能建造。E-mail：erintly@cqu.edu.cn
基金资助:
国家自然科学基金(52130801)

Dimension detection method for cable support curtain wall panels based on unbalanced optimal transport theory

TAN Liyun¹^,²(), LIU Jiepeng¹^,²(), LI Hantao³, ZENG Yan¹^,², LIAO Yue¹^,², WU Xiaofeng³, CUI Na¹^,²

¹ School of Civil Engineering, Chongqing University, Chongqing 400045, China
² Key Laboratory of New Technology for Construction of Cities in Mountain Area, Chongqing University, Chongqing 400045, China
³ Guangzhou Gezhouba Group Construction Engineering CO., LTD, , Guangzhou Guangdong 511466, China

Received:2024-11-11 Accepted:2025-04-25 Published:2025-10-30 Online:2025-09-10
First author：TAN Liyun (2001-), master student. Her main research interest covers intelligent construction. E-mail：erintly@cqu.edu.cn
Supported by:
National Natural Science Foundation of China(52130801)

摘要/Abstract

摘要：

玻璃幕墙因独特的美感和强大的塑型能力被广泛应用于大型场馆和地标建筑，其施工流程为先施工龙骨支撑系统再进行玻璃面板的安装与调试，但在具体操作时，建成龙骨的轴线通常与设计轴线有所偏差，导致后续玻璃面板施工困难。本文开展基于非平衡最优传输理论的幕墙点云轴线提取研究。该方法充分利用点云数据信息，结合非平衡最优传输理论，包括输入点云数据、点云数据预处理、随机采样获得初始轴线点集、提取杆件粗轴线和提取杆件精轴线等步骤，以此获得目标点云的轴线特征，后续再根据提取到的轴线获取幕墙尺寸。实验结果表明，该方法能有效提取幕墙龙骨的轴线，且具有良好的中心性与鲁棒性。通过与其他算法进行比较，证明了其有效性。计算得到的幕墙尺寸与实际测得的结果相比，误差在±2 mm内，在允许误差范围内。

关键词: 点云数据, 轴线提取, 非平衡最优传输, 玻璃幕墙, 智能建造

Abstract:

Glass curtain walls are widely used in large venues and landmark buildings due to their unique aesthetic appeal and powerful shaping capabilities. The construction process is to first construct the keel support system and then install and debug the glass panels. However, during the specific operation, the axis of the built keel usually deviates from the design axis, complicating subsequent construction of glass panels. Our research on axis extraction of curtain-wall based on unbalanced optimal transport theory was conducted. This method fully utilized point-cloud data information and combined non-equilibrium optimal transmission theory. Steps included inputting point cloud data, preprocessing point cloud data, random sampling to obtain an initial axis point set, extracting a thick axis of the rod, and extracting a fine axis of the rod. In this way, the axis features of the target point cloud were obtained. Then, the curtain-wall dimensions were obtained based on the extracted axis. Experimental results showed that this method can effectively extract the axis of the curtain-wall keel and exhibited strong centrality and robustness. The effectiveness of the method was demonstrated by comparison with other algorithms. Compared with the actual measured results, the calculated curtain-wall dimensions deviated by within ±2 mm, remaining within the allowable error range.

Key words: point cloud data, axis extraction, unbalanced optimal transport, glass curtain wall, intelligent construction

中图分类号:

TU767

谭立云, 刘界鹏, 李汉涛, 曾焱, 廖岳, 吴小峰, 崔娜. 基于非平衡最优传输理论的拉索式幕墙面板尺寸检测方法[J]. 图学学报, 2025, 46(5): 1134-1143.

TAN Liyun, LIU Jiepeng, LI Hantao, ZENG Yan, LIAO Yue, WU Xiaofeng, CUI Na. Dimension detection method for cable support curtain wall panels based on unbalanced optimal transport theory[J]. Journal of Graphics, 2025, 46(5): 1134-1143.

图/表 23

参考文献 23

[1]	FENG R Q, WU Y, SHEN S Z. Working mechanism of single-layer cable net supported glass curtain walls[J]. Advances in Structural Engineering, 2007, 10(2): 183-195.
[2]	YU Y, LIU T, ZHANG Q L, et al. Wind‐induced response of an L‐shaped cable support glass curtain wall[J]. Shock and Vibration, 2017, 2017: 4163045.
[3]	WANG J H, YI T Q, LIANG X, et al. Application of 3D laser scanning technology using laser radar system to error analysis in the curtain wall construction[J]. Remote Sensing, 2022, 15(1): 64.
[4]	PARK J S, PARK H S. High-precision dimensional measurement of a curtain wall cross-section using image super-resolution[J]. Materials Research Proceedings, 2023, 27: 223-227.
[5]	WU D C, LI Y, LI R, et al. Automatic curtain wall frame detection based on deep learning and cross-modal feature fusion[J]. Automation in Construction, 2024, 160: 105305.
[6]	TAGLIASACCHI A, ZHANG H, COHEN-OR D. Curve skeleton extraction from incomplete point cloud[J]. ACM Transactions on Graphics (TOG), 2009, 28(3): 71.
[7]	CAO J J, TAGLIASACCHI A, OLSON M, et al. Point cloud skeletons via Laplacian based contraction[C]// 2010 Shape Modeling International Conference. New York: IEEE Press, 2010: 187-197.
[8]	HUANG H, WU S H, COHEN-OR D, et al. L₁-medial skeleton of point cloud[J]. ACM Transactions on Graphics (TOG), 2013, 32(4): 65.
[9]	QIN H X, HAN J, LI N, et al. Mass-driven topology-aware curve skeleton extraction from incomplete point clouds[J]. IEEE Transactions on Visualization and Computer Graphics, 2020, 26(9): 2805-2817.
[10]	VILLANI C. Topics in optimal transportation[M]. Providence: American Mathematical Society, 2003: 370.
[11]	ZHANG J Y, ZHONG W X, MA P. A review on modern computational optimal transport methods with applications in biomedical research[M]// ZHAOY C, CHEND G. Modern Statistical Methods for Health Research. Cham: Springer, 2021: 279-300.
[12]	RUBNER Y, TOMASI C, GUIBAS L J. The earth mover’s distance as a metric for image retrieval[J]. International Journal of Computer Vision, 2000, 40(2): 99-121.
[13]	FRISCH U, MATARRESE S, MOHAYAEE R, et al. A reconstruction of the initial conditions of the universe by optimal mass transportation[J]. Nature, 2002, 417(6886): 260-262.
[14]	BENAMOU J D, BRENIER Y. Weak existence for the semigeostrophic equations formulated as a coupled Monge-Ampère/transport problem[J]. SIAM Journal on Applied Mathematics, 1998, 58(5): 1450-1461.
[15]	KANTOROVICH L. On the transfer of masses[J]. Doklady Akademii Nauk, 1942, 37(2): 227-229 (in Russian).
[16]	WASSERSTEIN L N. Markov processes over denumerable products of spaces describing large systems of automata[J]. Problemy Peredachi Informatsii, 1969, 5(3): 64-72.
[17]	PEYRÉ G, CUTURI M. Computational optimal transport[J]. Foundations and Trends® in Machine Learning, 2019, 11(5-6): 355-607.
[18]	CUTURI M. Sinkhorn distances: Lightspeed computation of optimal transport[C]// The 27th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2013: 2292-2300.
[19]	CAFFARELLI L A, McCann R J. Free boundaries in optimal transport and Monge-Ampere obstacle problems[J]. Annals of Mathematics, 2010, 171(2): 673-730.
[20]	CHIZAT L, PEYRÉ G, SCHMITZER B, et al. Scaling algorithms for unbalanced optimal transport problems[J]. Mathematics of Computation, 2018, 87(314): 2563-2609.
[21]	RUSU R B, BLODOW N, BEETZ M. Fast point feature histograms (FPFH) for 3D registration[C]// 2009 IEEE International Conference on Robotics and Automation. New York: IEEE Press, 2009: 3212-3217.
[22]	BESL P J, MCKAY N D. Method for registration of 3-D shapes[EB/OL]. [2024-07-11]. https://www.spiedigitallibrary.org/conference-proceedings-of-spie/1611/1/Method-for-registration-of-3-D-shapes/10.1117/12.57955.short.
[23]	中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会. GB/T 21086-2007 建筑幕墙[S]. 北京: 中国标准出版社, 2008.
	General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China, Standardization Administration of the People’s Republic of China. GB/T 21086-2007 Curtain wall for building[S]. Beijing: Standards Press of China, 2008 (in Chinese).

拉索玻璃编号	尺寸(左，上，右，下)	对角线(左上，左下)
LS-BL1-1	2 983.4, 2 005.9, 2 989.0, 2 002.5	3 602.2, 3 590.6
LS-BL1-2	2 989.0, 1 991.6, 2 994.5, 1 993.2	3 599.9, 3 589.1
LS-BL1-3	2 994.5, 1 994.5, 3 000.1, 1 995.7	3 606.8, 3 594.4
LS-BL1-4	3 000.1, 1 997.8, 3 005.6, 1 990.5	3 609.0, 3 600.4
LS-BL1-9	2 998.7, 2 019.2, 2 998.3, 2 012.0	3 619.8, 3 606.1
LS-BL1-10	2 998.3, 1 989.2, 2 998.0, 1 993.9	3 605.5, 3 593.2
LS-BL1-11	2 998.0, 2 002.1, 2 997.6, 1 999.0	3 610.6, 3 597.4
LS-BL1-12A	2 997.6, 1 948.9, 2 997.5, 1 980.0	3 598.0, 3 569.7
LS-BL1-16A	3 010.0, 1 943.8, 3 008.1, 1 947.8	3 594.0, 3 572.7
LS-BL1-34A	2 992.8, 1 940.0, 2 982.2, 1 943.8	3 579.0, 3 547.3
LS-BL1-52A	3 007.4, 1 936.1, 2 996.8, 1 940.0	3 598.0, 3 548.4
LS-BL1-70A	3 007.7, 1 932.4, 2 993.7, 1 936.1	3 605.1, 3 534.8

拉索玻璃编号	尺寸(左，上，右，下)	对角线(左上，左下)
LS-BL1-1	2 983.4, 2 005.9, 2 989.0, 2 002.5	3 602.2, 3 590.6
LS-BL1-2	2 989.0, 1 991.6, 2 994.5, 1 993.2	3 599.9, 3 589.1
LS-BL1-3	2 994.5, 1 994.5, 3 000.1, 1 995.7	3 606.8, 3 594.4
LS-BL1-4	3 000.1, 1 997.8, 3 005.6, 1 990.5	3 609.0, 3 600.4
LS-BL1-9	2 998.7, 2 019.2, 2 998.3, 2 012.0	3 619.8, 3 606.1
LS-BL1-10	2 998.3, 1 989.2, 2 998.0, 1 993.9	3 605.5, 3 593.2
LS-BL1-11	2 998.0, 2 002.1, 2 997.6, 1 999.0	3 610.6, 3 597.4
LS-BL1-12A	2 997.6, 1 948.9, 2 997.5, 1 980.0	3 598.0, 3 569.7
LS-BL1-16A	3 010.0, 1 943.8, 3 008.1, 1 947.8	3 594.0, 3 572.7
LS-BL1-34A	2 992.8, 1 940.0, 2 982.2, 1 943.8	3 579.0, 3 547.3
LS-BL1-52A	3 007.4, 1 936.1, 2 996.8, 1 940.0	3 598.0, 3 548.4
LS-BL1-70A	3 007.7, 1 932.4, 2 993.7, 1 936.1	3 605.1, 3 534.8

拉索玻璃编号	尺寸(左，上，右，下)	对角线(左上，左下)
LS-BL11	2 983.5, 2 005.5, 2 988.7, 2 002.9	3 603.8, 3 590.8
LS-BL1-2	2 988.7, 1 992.5, 2 994.7, 1 991.4	3 600.4, 3 589.3
LS-BL1-3	2 994.7, 1 993.5, 3 000.3, 1 995.3	3 605.2, 3 593.7
LS-BL1-4	3 000.3, 1 998.2, 3 005.9, 1 989.6	3 609.2, 3 600.8
LS-BL1-9	2 998.0, 2 020.3, 2 998.2, 2 010.9	3 620.2, 3 607.4
LS-BL1-10	2 998.2, 1 991.1, 2 998.6, 1 992.0	3 606.4, 3 593.1
LS-BL1-11	2 998.6, 2 002.5, 2 996.8, 1 998.6	3 610.2, 3 597.1
LS-BL1-12A	2 996.8, 1 948.9, 2 996.9, 1 981.2	3 597.8, 3 570.8
LS-BL1-16A	3 008.1, 1 944.1, 3 009.6, 1 948.9	3 595.4, 3 573.8
LS-BL1-34A	2 993.7, 1 938.9, 2 982.3, 1 944.1	3 580.3, 3 548.2
LS-BL1-52A	3 008.0, 1 936.4, 2 995.9, 1 938.9	3 596.4, 3 547.7
LS-BL1-70A	3 007.3, 1 931.4, 2 994.4, 1 936.4	3 604.7, 3 534.2

拉索玻璃编号	尺寸(左，上，右，下)	对角线(左上，左下)
LS-BL11	2 983.5, 2 005.5, 2 988.7, 2 002.9	3 603.8, 3 590.8
LS-BL1-2	2 988.7, 1 992.5, 2 994.7, 1 991.4	3 600.4, 3 589.3
LS-BL1-3	2 994.7, 1 993.5, 3 000.3, 1 995.3	3 605.2, 3 593.7
LS-BL1-4	3 000.3, 1 998.2, 3 005.9, 1 989.6	3 609.2, 3 600.8
LS-BL1-9	2 998.0, 2 020.3, 2 998.2, 2 010.9	3 620.2, 3 607.4
LS-BL1-10	2 998.2, 1 991.1, 2 998.6, 1 992.0	3 606.4, 3 593.1
LS-BL1-11	2 998.6, 2 002.5, 2 996.8, 1 998.6	3 610.2, 3 597.1
LS-BL1-12A	2 996.8, 1 948.9, 2 996.9, 1 981.2	3 597.8, 3 570.8
LS-BL1-16A	3 008.1, 1 944.1, 3 009.6, 1 948.9	3 595.4, 3 573.8
LS-BL1-34A	2 993.7, 1 938.9, 2 982.3, 1 944.1	3 580.3, 3 548.2
LS-BL1-52A	3 008.0, 1 936.4, 2 995.9, 1 938.9	3 596.4, 3 547.7
LS-BL1-70A	3 007.3, 1 931.4, 2 994.4, 1 936.4	3 604.7, 3 534.2

拉索玻璃编号	左	上	右	下	左上	左下
LS-BL1-1	−0.1	0.4	0.3	−0.4	−1.6	−0.2
LS-BL1-2	0.3	−0.9	−0.2	1.8	−0.5	−0.2
LS-BL1-3	−0.2	1	−0.2	0.4	1.6	0.7
LS-BL1-4	−0.2	−0.4	−0.3	0.9	−0.2	−0.4
LS-BL1-9	0.7	−1.1	0.1	1.1	−0.4	−1.3
LS-BL1-10	0.1	−1.9	−0.6	1.9	−0.9	0.1
LS-BL1-11	−0.6	−0.4	0.8	0.4	0.4	0.3
LS-BL1-12A	0.8	0	0.6	−1.2	0.2	−1.1
LS-BL1-16A	1.9	−0.3	−1.5	−1.1	−1.4	−1.1
LS-BL1-34A	−0.9	1.1	−0.1	−0.3	−1.3	−0.9
LS-BL1-52A	−0.6	−0.3	0.9	1.1	1.6	0.7
LS-BL1-70A	0.4	1	−0.7	−0.3	0.4	0.6

基于非平衡最优传输理论的拉索式幕墙面板尺寸检测方法

Dimension detection method for cable support curtain wall panels based on unbalanced optimal transport theory

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 23

参考文献 23

相关文章 7

编辑推荐

Metrics

本文评价

拉索玻璃编号	尺寸(左、上、右、下)
拉索玻璃编号	本文算法	ROSA法	MDCS法
1	0.013 6, 0.007 2, 0.009 9, 0.007 8	0.013 2, 0.042 2, 0.013 8, 0.014 2	0.059 5, -, 0.040 3, -
2	0.009 9, 0.008 1 , 0.006 7, 0.005 9	0.013 8, 0.014 2, 0.012 6, 0.034 5	0.040 3, 0.036 7, 0.171 6
3	0.006 7, 0.012 1, 0.016 8, 0.007 7	0.012 6, 0.012 2, 0.013 2, 0.013 5	0.036 7, -, 0.025 3, -
4	0.019 9, 0.007 8, 0.017 2, 0.016 3	0.031 4, 0.014 2, 0.016 5, 0.031 1	0.051 5, -, 0.107 6, 0.073 6
5	0.017 2, 0.005 9, 0.008 2, 0.008 4	0.016 5, 0.034 5, 0.011 9, 0.007 2	0.107 6, 0.171 6, 0.074 8, 0.133 4
6	0.008 2, 0.007 7, 0.005 6, 0.006 2	0.011 9, 0.013 5, 0.008 6, 0.007 4	0.074 8, -, 0.076 2, -

拉索玻璃编号	尺寸(左，上，右，下)	对角线(左上，左下)
1	3 005.5, 2 000.6, 3 002.6, 2 000.6	3 618.6, 3 599.9
2	3 002.6, 2 013.7, 2 999.7, 2 013.6	3 623.5, 3 604.7
3	2 999.7, 1 996.4, 2 996.8, 1 993.7	3 609.9, 3 592.8
4	2 993.3, 2 000.6, 2 991.4, 2 000.6	3 606.9, 3 592.2
5	2 991.4, 2 013.6, 2 989.4, 2 013.5	3 612.5, 3 597.8
6	2 989.4, 1 993.7, 2 987.5, 1 991.1	3 598.3, 3 585.2

[1]	方程浩, 王康侃. 基于半监督学习的单视角点云三维人体姿态与形状估计[J]. 图学学报, 2025, 46(2): 393-401.
[2]	黄祥, 王红星, 顾徐, 孟悦, 王浩羽. 一种新的基于特殊离群样本优化的三维点云特征选择算法[J]. 图学学报, 2022, 43(5): 884-891.
[3]	刘玉珍, 李楠, 陶志勇. 基于环查询和通道注意力的点云分类与分割[J]. 图学学报, 2022, 43(4): 616-623.
[4]	郑炜 , 林佳瑞 , 杨程 , 闫克霄 , 程宇 . 工程监理人员绩效智能评价方法研究综述 [J]. 图学学报, 2021, 42(6): 1002-1010.
[5]	石力文 1,2,3，解琳琳 1,4，侯妙乐 1,2,3，李爱群 1,4，胡云岗 1,2,3，刘浩宇 1,2,3 . 基于显著几何特征的古木建筑关键构件多 LoD 尺寸信息提取方法[J]. 图学学报, 2019, 40(4): 651-658.
[6]	闫胜昝，章健，苏本跃. 不规则复杂牙冠模型的建立及优化分析[J]. 图学学报, 2015, 36(2): 198-204.
[7]	吴雪梅，文珈，于广滨，李瑰贤，单德彬. 基于三坐标测量数据的点云曲面重构[J]. 图学学报, 2011, 32(2): 68-72.