Hydrant Segmentation and Extraction Using Deep Learning Models for Large-Scale Urban Point Clouds
Publication: Construction Research Congress 2024
ABSTRACT
This study addresses the gap in research concerning the segmentation of smaller infrastructure, such as fire hydrants, within large-scale point cloud models. The RandLA-Net model, an efficient and robust deep learning algorithm known for its ability to process vast quantities of data while maintaining high accuracy rapidly, was evaluated for this task. A comprehensive dataset was constructed using LiDAR technology to gather point cloud data from diverse urban scenes. The data was manually segmented to assign unique class identifiers to distinct objects within the point cloud. The model was trained using these annotated data, and its performance was assessed using Intersection over Union (IoU), a typical evaluation metric for segmentation tasks. The results demonstrated the model’s impressive capability in segmenting different classes within large-scale urban point cloud data. However, segmenting smaller, less distinctive objects like fire hydrants revealed room for improvement. Future work should focus on improving the recognition of complex or underrepresented classes and enhancing training stability. The study’s findings provide critical insights into the capabilities and limitations of the RandLA-Net model in the context of large-scale, real-world point cloud data segmentation tasks.
Get full access to this article
View all available purchase options and get full access to this chapter.
REFERENCES
Bonczak, B., and C. E. Kontokosta. 2019. “Large-scale parameterization of 3D building morphology in complex urban landscapes using aerial LiDAR and city administrative data.” Computers, Environment, and Urban Systems, 73: 126–142. https://doi.org/10.1016/j.compenvurbsys.2018.09.004.
Chen, J., Z. Kira, and Y. K. Cho. 2019. “Deep Learning Approach to Point Cloud Scene Understanding for Automated Scan to 3D Reconstruction.” Journal of Computing in Civil Engineering, 33 (4): 04019027. American Society of Civil Engineers. https://doi.org/10.1061/(ASCE)CP.1943-5487.0000842.
Chen, M., A. Feng, R. McAlinden, and L. Soibelman. 2020. “Photogrammetric Point Cloud Segmentation and Object Information Extraction for Creating Virtual Environments and Simulations.” Journal of Management in Engineering, 36 (2): 04019046. American Society of Civil Engineers. https://doi.org/10.1061/(ASCE)ME.1943-5479.0000737.
Chew, A. W. Z., A. Ji, and L. Zhang. 2022. “Large-scale 3D point-cloud semantic segmentation of urban and rural scenes using data volume decomposition coupled with pipeline parallelism.” Automation in Construction, 133: 103995. https://doi.org/10.1016/j.autcon.2021.103995.
“CloudCompare - Open Source project.” n.d. Accessed May 15, 2023. https://www.danielgm.net/cc/.
Deng, T., K. Zhang, and Z.-J. M. Shen. 2021. “A systematic review of a digital twin city: A new pattern of urban governance toward smart cities.” Journal of Management Science and Engineering, 6 (2): 125–134. https://doi.org/10.1016/j.jmse.2021.03.003.
Geng, Y., Z. Wang, L. Jia, Y. Qin, Y. Chai, K. Liu, and L. Tong. 2023. “3DGraphSeg: A Unified Graph Representation-Based Point Cloud Segmentation Framework for Full-Range Highspeed Railway Environments.” IEEE Transactions on Industrial Informatics, 1–13. https://doi.org/10.1109/TII.2023.3246492.
Gong, J. 2014. “A Remote Sensing-based Approach for Assessing and Visualizing Post-Sandy Damage and Resiliency Rebuilding Needs.” 1259–1268. American Society of Civil Engineers. https://doi.org/10.1061/9780784413517.129.
Hernández-García, D.-E., J.-J. Gonzalez-Barbosa, J.-B. Hurtado-Ramos, F.-J. Ornelas-Rodríguez, E. Castillo Castaneda, A. Ramírez, A. I. García, R. Gonzalez-Barbosa, and J. G. Aviña-Cervantez. 2011. “3D city models: Mapping approach using LIDAR technology.” CONIELECOMP 2011, 21st International Conference on Electrical Communications and Computers, 206–211.
Hu, Q., B. Yang, L. Xie, S. Rosa, Y. Guo, Z. Wang, N. Trigoni, and A. Markham. 2020. “RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds.”.
Ji, A., A. W. Z. Chew, X. Xue, and L. Zhang. 2022. “An encoder-decoder deep learning method for multi-class object segmentation from 3D tunnel point clouds.” Automation in Construction, 137: 104187. https://doi.org/10.1016/j.autcon.2022.104187.
Liu, C., D. Zeng, A. Akbar, H. Wu, S. Jia, Z. Xu, and H. Yue. 2022. “Context-Aware Network for Semantic Segmentation Toward Large-Scale Point Clouds in Urban Environments.” IEEE Transactions on Geoscience and Remote Sensing, 60: 1–15. https://doi.org/10.1109/TGRS.2022.3182776.
Mirzaei, K., M. Arashpour, E. Asadi, H. Masoumi, Y. Bai, and A. Behnood. 2022. “3D point cloud data processing with machine learning for construction and infrastructure applications: A comprehensive review.” Advanced Engineering Informatics, 51: 101501. https://doi.org/10.1016/j.aei.2021.101501.
Park, Y., and J.-M. Guldmann. 2019. “Creating 3D city models with building footprints and LIDAR point cloud classification: A machine learning approach.” Computers, Environment, and Urban Systems, 75: 76–89. https://doi.org/10.1016/j.compenvurbsys.2019.01.004.
Richter, R., M. Behrens, and J. Döllner. 2013. “Object class segmentation of massive 3D point clouds of urban areas using point cloud topology.” International Journal of Remote Sensing, 34 (23): 8408–8424. Taylor & Francis. https://doi.org/10.1080/01431161.2013.838710.
Information & Authors
Information
Published In
History
Published online: Mar 18, 2024
ASCE Technical Topics:
- Artificial intelligence and machine learning
- Computer networks
- Computer programming
- Computing in civil engineering
- Continuum mechanics
- Disaster risk management
- Disasters and hazards
- Dynamics (solid mechanics)
- Engineering fundamentals
- Engineering mechanics
- Fires
- Infrastructure
- Man-made disasters
- Model accuracy
- Models (by type)
- Motion (dynamics)
- Neural networks
- Solid mechanics
- Uncertainty principles
- Urban and regional development
- Urban areas
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.