Smart City Gnosys

Smart city article details

Title	Self-Supervised Pretraining Framework For Extracting Global Structures From Building Point Clouds Via Completion
ID_Doc	48202
Authors	Yang H.; Wang R.
Year	2024
Published	IEEE Transactions on Geoscience and Remote Sensing, 62
DOI	http://dx.doi.org/10.1109/TGRS.2024.3477423
Abstract	The exterior structural information of buildings are crucial for advancing smart city initiatives and reconstructing 3-D edifices. However, practical obstacles, such as sparse or incomplete building point clouds - stemming from various scanning angles or sensor limitations - present significant challenges. To mitigate the high costs and labor demands associated with data labeling, we introduce an innovative pretraining framework with self-supervised learning (SSL) that incorporates a modified point cloud completion (PCC) subnetwork to extract building structures. Specifically, the modified PCC subnetwork completes the original partial building point clouds by capturing both fine-grained and high-level semantic information of 3-D shapes. Following this, self-supervised feature extractor using a masked autoencoder (MAE) and a multiscale feature mechanism generates pointwise features from the completed building point clouds. We evaluate the effectiveness of the proposed integrated framework in extracting global structures using both established wireframe construction methods and our newly proposed edge point identification that incorporates a novel edge point regression loss. Extensive experimental results demonstrate that our modified PCC network reaches a 93.5% convergence rate that is higher than the results from competing methods. Our self-supervised pretraining framework extracts more accurate global structures with better loss convergence than traditional edge point identification loss designs. Finally, our combined framework improves performance for subsequent processes (such as wireframe construction and edge point identification) when using completed datasets instead of the original partial datasets. © 1980-2012 IEEE.
Author Keywords	Building point cloud completion (PCC) subnetwork; edge point identification; light detection and ranging (LiDAR) point cloud; self-supervised learning (SSL)

Similar Articles

Id	Similarity	Authors	Title	Published
4556	0.942	Yang H.; Xu S.; Xu S.	A Self-Supervised Pretraining Framework For Context-Aware Building Edge Extraction From 3-D Point Clouds	IEEE Geoscience and Remote Sensing Letters, 22 (2025)
13123	0.88	Wang R.; Huang S.; Yang H.	Building3D: An Urban-Scale Dataset And Benchmarks For Learning Roof Structures From Point Clouds	Proceedings of the IEEE International Conference on Computer Vision (2023)
40412	0.876	Li S.; Yan F.; Ma K.; Hu Q.; Wang F.; Liu W.	Optimal Feature-Guided Position-Shape Dual Optimization For Building Point Cloud Facade Detail Enhancement	Remote Sensing, 16, 22 (2024)
25925	0.867	Huang Y.; Zang Y.; Jiang Q.; Mi W.	Extraction Of Building Outlines From Airborne Lidar Point Clouds Using Line-Cnn Based On Deep Network; [采用 Line-Cnn 深度学习网络的机载点云建筑轮廓线提取]	Journal of Geo-Information Science, 26, 9 (2024)
1779	0.863	Zuo Z.; Li Y.	A Framework For Reconstructing Building Parametric Models With Hierarchical Relationships From Point Clouds	International Journal of Applied Earth Observation and Geoinformation, 119 (2023)
13030	0.861	Zhang Y.; Wang T.; Lin X.; Zhao Z.; Wang X.	Building Extraction From Lidar Point Clouds Based On Revised Randla-Net	International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives, 48, 1 (2024)
102	0.861	Yu D.; Ji S.; Wei S.; Khoshelham K.	3-D Building Instance Extraction From High-Resolution Remote Sensing Images And Dsm With An End-To-End Deep Neural Network	IEEE Transactions on Geoscience and Remote Sensing, 62 (2024)
15252	0.858	Zhang W.; Chen J.; Tan G.	Complex Roof Structure Reconstruction By 3D Primitive Fitting From Point Clouds; [基于 3D 基元拟合的复杂屋顶点云三维重建]	Journal of Geo-Information Science, 25, 8 (2023)