Automated Semantic Segmentation of Indoor Point Clouds from Close-Range Images with Three-Dimensional Deep Learning
Abstract
:1. Introduction
2. Related Work
2.1. Three-Dimensional Point Cloud Classification
2.2. Semantic Segmentation and Modeling
3. Methodology
3.1. Three-Dimensional Point Cloud Classification
3.1.1. Sample Data
- S3DIS Dataset
- 2.
- Close-range Images
3.1.2. Sample Training
3.2. Semantic Segmentation and Modeling
3.2.1. Category Extraction
3.2.2. Labeled Category
3.2.3. Removal of Outliers
3.2.4. Feature Extraction
3.2.5. Three-Dimensional Modeling
4. Experiments and Analyses
4.1. Test Area: Civil Engineering Building
4.2. Three-Dimensional Point Cloud Classification
4.2.1. S3DIS Dataset
4.2.2. Civil Engineering Building
4.2.3. Discussion of Classification Results
- Number of point clouds
- 2.
- Geometric distribution of point cloud
4.3. Semantic Segmentation and Modeling
4.3.1. S3DIS Modeling
4.3.2. 2F Corridor of Civil Engineering Building
4.4. Evaluation of 3D Model
5. Conclusions and Future Work
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Griffiths, D.; Boehm, J. A Review on Deep Learning Techniques for 3D Sensed Data Classification. Remote Sens. 2019, 11, 1499. [Google Scholar] [CrossRef]
- Wang, Y.; Ahsan, U.; Li, H.; Hagen, M. A Comprehensive Review of Modern Object Segmentation Approaches. Found. Trends Comput. Graph. Vis. 2022, 13, 111–283. [Google Scholar] [CrossRef]
- Guo, Y.-C.; Weng, T.-H.; Fischer, R.; Fu, L.-C. 3D semantic segmentation based on spatial-aware convolution and shape completion for augmented reality applications. Comput. Vis. Image Underst. 2022, 224, 103550. [Google Scholar] [CrossRef]
- Chen, H.; Liang, M.; Liu, W.; Wang, W.; Liu, P.X. An approach to boundary detection for 3D point clouds based on DBSCAN clustering. Pattern Recognit. 2022, 124, 108431. [Google Scholar] [CrossRef]
- Biagini, C.; Capone, P.; Donato, V.; Facchini, N. Towards the BIM implementation for historical building restoration sites. Autom. Constr. 2016, 71, 74–86. [Google Scholar] [CrossRef]
- Bello, S.A.; Yu, S.; Wang, C.; Adam, J.M.; Li, J. Review: Deep Learning on 3D Point Clouds. Remote. Sens. 2020, 12, 1729. [Google Scholar] [CrossRef]
- Camuffo, E.; Mari, D.; Milani, S. Recent Advancements in Learning Algorithms for Point Clouds: An Updated Overview. Sensors 2022, 22, 1357. [Google Scholar] [CrossRef]
- Diab, A.; Kashef, R.; Shaker, A. Deep Learning for LiDAR Point Cloud Classification in Remote Sensing. Sensors 2022, 22, 7868. [Google Scholar] [CrossRef]
- Sun, Y.; Paik, J.; Koschan, A.; Abidi, M. 3D reconstruction of indoor and outdoor scenes using a mobile range scanner. In Proceedings of the 2002 International Conference on Pattern Recognition, IEEE, Quebec City, QC, Canada, 11–15 August 2002. [Google Scholar]
- Wu, H.; Yue, H.; Han, J.; Zhang, S. Indoor Map Boundary Correction Based on Normalized Total Least Squares of Condition Equation. J. Surv. Eng. 2021, 147, 04021015. [Google Scholar] [CrossRef]
- Armeni, I.; Sener, O.; Zamir, A.R.; Jiang, H.; Brilakis, I.; Fischer, M.; Savarese, S. 3D semantic parsing of large-scale indoor spaces. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016. [Google Scholar]
- Wang, T.; Wang, Q.; Ai, H.; Zhang, L. Semantics-and-Primitives-Guided Indoor 3D Reconstruction from Point Clouds. Remote. Sens. 2022, 14, 4820. [Google Scholar] [CrossRef]
- Wang, Y.; Sun, Y.; Liu, Z.; Sarma, S.E.; Bronstein, M.M.; Solomon, J.M. Dynamic Graph CNN for Learning on Point Clouds. ACM Trans. Graph. 2019, 38, 1–12. [Google Scholar] [CrossRef]
- Guo, Y.; Wang, H.; Hu, Q.; Liu, H.; Liu, L.; Bennamoun, M. Deep Learning for 3D Point Clouds: A Survey. IEEE Trans. Pattern Anal. Mach. Intell. 2021, 43, 4338–4364. [Google Scholar] [CrossRef]
- Liu, W.; Sun, J.; Li, W.; Hu, T.; Wang, P. Deep Learning on Point Clouds and Its Application: A Survey. Sensors 2019, 19, 4188. [Google Scholar] [CrossRef]
- Yang, J.; Zou, B.; Qiu, H.; Li, Z. MLFNet-Point Cloud Semantic Segmentation Convolution Network Based on Multi-Scale Feature Fusion. IEEE Access 2021, 9, 44950–44962. [Google Scholar] [CrossRef]
- Chen, P.; Luo, Z.; Shi, W. Hysteretic mapping and corridor semantic modeling using mobile LiDAR systems. ISPRS J. Photogramm. Remote. Sens. 2022, 186, 267–284. [Google Scholar] [CrossRef]
- Ochmann, S.; Vock, R.; Klein, R. Automatic reconstruction of fully volumetric 3D building models from oriented point clouds. ISPRS J. Photogramm. Remote. Sens. 2019, 151, 251–262. [Google Scholar] [CrossRef]
- Xie, Y.; Tian, J.; Zhu, X.X. Linking Points with Labels in 3D: A Review of Point Cloud Semantic Segmentation. IEEE Geosci. Remote Sens. Mag. 2020, 8, 38–59. [Google Scholar] [CrossRef]
- Rashdi, R.; Martínez-Sánchez, J.; Arias, P.; Qiu, Z. Scanning Technologies to Building Information Modelling: A Review. Infrastructures 2022, 7, 49. [Google Scholar] [CrossRef]
- Jhaldiyal, A.; Chaudhary, N. Semantic segmentation of 3D LiDAR data using deep learning: A review of projection-based methods. Appl. Intell. 2022, 1–12. [Google Scholar] [CrossRef]
- Lin, H.; Wu, S.; Chen, Y.; Li, W.; Luo, Z.; Guo, Y.; Wang, C.; Li, J. Semantic segmentation of 3D indoor LiDAR point clouds through feature pyramid architecture search. ISPRS J. Photogramm. Remote. Sens. 2021, 177, 279–290. [Google Scholar] [CrossRef]
- Jeon, J.; Jung, J.; Kim, J.; Lee, S. Semantic Reconstruction: Reconstruction of Semantically Segmented 3D Meshes via Volumetric Semantic Fusion. Comput. Graph. Forum 2018, 37, 25–35. [Google Scholar] [CrossRef]
- Qi, C.R.; Su, H.; Mo, K.; Guibas, L.J. PointNet: Deep learning on point sets for 3D classification and segmentation. In Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, Honolulu, HI, USA, 21–26 July 2017. [Google Scholar]
- Qi, C.R.; Yi, L.; Su, H.; Guibas, L.J. PointNet++: Deep hierarchical feature learning on point sets in a metric space. Adv. Neural Inf. Process. Syst. 2017, 30, 5105–5114. [Google Scholar]
- Mitropoulou, A.; Georgopoulos, A. An Automated Process to Detect Edges in Unorganized Point Clouds. ISPRS Ann. Photogramm. Remote. Sens. Spat. Inf. Sci. 2019, IV-2/W6, 99–105. [Google Scholar] [CrossRef]
- Kong, G.; Fan, H.; Lobaccaro, G. Automatic building outline extraction from ALS point cloud data using generative adversarial network. Geocarto Int. 2022, 1–18. [Google Scholar] [CrossRef]
- Mahphood, A.; Arefi, H. Grid-based building outline extraction from ready-made building points. Autom. Constr. 2022, 139, 104321. [Google Scholar] [CrossRef]
- Chen, X.; Yu, K. Feature Line Generation and Regularization from Point Clouds. IEEE Trans. Geosci. Remote Sens. 2019, 57, 9779–9790. [Google Scholar] [CrossRef]
- Awrangjeb, M. Using point cloud data to identify, trace, and regularize the outlines of buildings. Int. J. Remote Sens. 2016, 37, 551–579. [Google Scholar] [CrossRef]
- Wang, Y.; Jeschke, S. Edge extraction by merging 3D point cloud and 2D image data. In Proceedings of the 2013 10th International Conference and Expo on Emerging Technologies for a Smarter World (CEWIT), IEEE, Melville, NY, USA, 21–22 October 2013. [Google Scholar]
- Borges, P.; Zlot, R.; Bosse, M.; Nushke, S.; Tews, A. Vision-based localization using an edge map extracted from 3D laser range data. In Proceedings of the 2010 IEEE International Conference on Robotics and Automation, IEEE, Anchorage, AK, USA, 3–7 May 2010. [Google Scholar]
- Sampath, A.; Shan, J. Building Boundary Tracing and Regularization from Airborne Lidar Point Clouds. Photogramm. Eng. Remote Sens. 2007, 73, 805–812. [Google Scholar] [CrossRef]
- Jiang, S.; Jiang, C.; Jiang, W. Efficient structure from motion for large-scale UAV images: A review and a comparison of SfM tools. ISPRS J. Photogramm. Remote Sens. 2020, 167, 230–251. [Google Scholar] [CrossRef]
- Braun, A.; Tuttas, S.; Borrmann, A.; Stilla, U. Improving progress monitoring by fusing point clouds, semantic data and computer vision. Autom. Constr. 2020, 116, 103210. [Google Scholar] [CrossRef]
- Torr, P.H.S.; Zisserman, A. MLESAC: A New Robust Estimator with Application to Estimating Image Geometry. Comput. Vis. Image Underst. 2000, 78, 138–156. [Google Scholar] [CrossRef] [Green Version]
- Macher, H.; Landes, T.; Grussenmeyer, P. From Point Clouds to Building Information Models: 3D Semi-Automatic Reconstruction of Indoors of Existing Buildings. Appl. Sci. 2017, 7, 1030. [Google Scholar] [CrossRef] [Green Version]
Conference_Room2 | Predicted Class | ||||||
---|---|---|---|---|---|---|---|
Beam | Ceiling | Column | Floor | Wall | % | ||
True Class | beam | 8214 | 287 | 3806 | 66.7 | ||
ceiling | 99 | 61,073 | 24 | 56 | 99.7 | ||
column | 276 | 7 | 3704 | 4084 | 45.9 | ||
floor | 17 | 425 | 47,700 | 1204 | 96.7 | ||
wall | 2627 | 9476 | 152 | 28,801 | 70.2 | ||
% | 95.6 | 95.4 | 27.2 | 99.7 | 75.9 | 86.9 |
Office2 | Predicted Class | ||||||
---|---|---|---|---|---|---|---|
Beam | Ceiling | Column | Floor | Wall | % | ||
True Class | beam | 5509 | 10 | 28 | 99.3 | ||
ceiling | 44 | 21,155 | 3 | 8 | 99.7 | ||
column | 22 | 1179 | 3 | 97.9 | |||
floor | 9 | 15,211 | 130 | 99.1 | |||
wall | 61 | 356 | 1177 | 3 | 28,820 | 94.7 | |
% | 98.1 | 98.2 | 49.8 | 100.0 | 99.4 | 97.5 |
Office6 | Predicted Class | ||||||
---|---|---|---|---|---|---|---|
Beam | Ceiling | Column | Floor | Wall | % | ||
True Class | beam | 4125 | 2693 | 130 | 59.4 | ||
ceiling | 19 | 16,823 | 2 | 47 | 99.6 | ||
column | 1176 | 5 | 99.6 | ||||
floor | 14,390 | 353 | 97.6 | ||||
wall | 23 | 97 | 1235 | 22 | 20,300 | 93.6 | |
% | 99.0 | 85.8 | 48.7 | 99.8 | 97.4 | 92.5 |
Ground Truth | Segmented Results | |
---|---|---|
Conference _Room2 | ||
Office_2 | ||
Office_6 |
Corridor 2F | Predicted Class | ||||||
---|---|---|---|---|---|---|---|
Beam | Ceiling | Column | Floor | Wall | % | ||
True Class | beam | 4642 | 51 | 538 | 818 | 76.7 | |
ceiling | 93 | 23,378 | 43 | 71 | 99.1 | ||
column | 8 | 2 | 5972 | 398 | 93.6 | ||
floor | 80 | 41,111 | 1345 | 96.6 | |||
wall | 585 | 1528 | 2300 | 52,205 | 92.2 | ||
% | 87.1 | 93.7 | 66.9 | 100 | 95.2 | 94.2 |
Ground Truth | Segmented Result | |
---|---|---|
Corridor 2F of Civil Engineering Building |
Line | Model Length | Actual Length | Difference | Line | Model Length | Actual Length | Difference |
---|---|---|---|---|---|---|---|
AB | 0.094 | 0.09 | 0.004 | BD | 0.259 | 0.232 | −0.027 |
BC | 1.715 | 1.721 | −0.006 | EF | 0.286 | 0.269 | 0.017 |
DE | 1.715 | 1.721 | −0.006 | HJ | 0.347 | 0.321 | 0.026 |
FG | 0.337 | 0.322 | 0.015 | IK | 0.347 | 0.324 | 0.023 |
HI | 2.032 | 2.05 | −0.018 | GI | 4.671 | 4.643 | 0.028 |
JK | 2.032 | 2.052 | −0.02 | NM | 0.360 | 0.304 | 0.056 |
LM | 1.778 | 1.784 | −0.006 | KO | 4.631 | 4.716 | −0.085 |
NO | 0.334 | 0.325 | 0.009 | FP | 0.749 | 0.781 | −0.032 |
AL | 10.598 | 10.609 | 0.011 | FQ | 3.455 | 3.430 | 0.025 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Hsieh, C.-S.; Ruan, X.-J. Automated Semantic Segmentation of Indoor Point Clouds from Close-Range Images with Three-Dimensional Deep Learning. Buildings 2023, 13, 468. https://doi.org/10.3390/buildings13020468
Hsieh C-S, Ruan X-J. Automated Semantic Segmentation of Indoor Point Clouds from Close-Range Images with Three-Dimensional Deep Learning. Buildings. 2023; 13(2):468. https://doi.org/10.3390/buildings13020468
Chicago/Turabian StyleHsieh, Chia-Sheng, and Xiang-Jie Ruan. 2023. "Automated Semantic Segmentation of Indoor Point Clouds from Close-Range Images with Three-Dimensional Deep Learning" Buildings 13, no. 2: 468. https://doi.org/10.3390/buildings13020468