DSC-Net: learning discriminative spatial contextual features for semantic segmentation of large-scale ancient architecture point clouds
-
Published:2024-08-02
Issue:1
Volume:12
Page:
-
ISSN:2050-7445
-
Container-title:Heritage Science
-
language:en
-
Short-container-title:Herit Sci
Author:
Zhao Jianghong,Liu Rui,Hua Xinnan,Yu Haiquan,Zhao Jifu,Wang Xin,Yang Jia
Abstract
AbstractSemantic segmentation of point cloud data of architectural cultural heritage is of significant importance for HBIM modeling, disease extraction and analysis, and heritage restoration research fields. In the semantic segmentation task of architectural point cloud data, especially for the protection and analysis of architectural cultural heritage, the previous deep learning methods have poor segmentation effects due to the complexity and unevenness of the data, the high geometric feature similarity between different components, and the large scale changes. To this end, this paper proposes a novel encoder-decoder architecture called DSC-Net. It consists of an encoder-decoder structure based on point random sampling and several fully connected layers for semantic segmentation. To overcome the loss of key features caused by random downsampling, DSC-Net has developed two new feature aggregation schemes: the enhanced dual attention pooling module and the global context feature module, to learn discriminative features for the challenging scenes mentioned above. The former fully considers the topology and semantic similarity of neighboring points, generating attention features that can distinguish categories with similar structures. The latter uses spatial location and neighboring volume ratio to provide an overall view of different types of architectural scenes, helping the network understand the spatial relationships and hierarchical structures between different architectural elements. The proposed modules can be easily embedded into various network architectures for point cloud semantic segmentation. We conducted experiments on multiple datasets, including the ancient architecture dataset, the ArCH architectural cultural heritage dataset, and the publicly available architectural segmentation dataset S3DIS. The results show that the mIoU reached 63.56%, 55.84%, and 71.03% respectively. The experimental results prove that our method has the best segmentation effect in dealing with challenging architectural cultural heritage data and also demonstrates its practicality in a wider range of architectural point cloud segmentation applications.
Funder
National Key R&D Program Project
Open Fund Project of State Key Laboratory of Geographic Information Engineering
Open Fund of State Key Laboratory of Surveying, Mapping and Remote Sensing Information Engineering of Wuhan University
Open Research Fund Project of the Key Laboratory of Digital Mapping and Land Information Application of the Ministry of Natural Resources
Software Science Research Project of the Ministry of Housing and Urban Rural Development
Beijing Social Science Foundation Decision Consulting Major Project
National Natural Science Foundation of China
Publisher
Springer Science and Business Media LLC
Reference39 articles.
1. Zhang R, Zhou X, Zhao J, Min C. A semantic segmentation algorithm of ancient building’s point cloud data. Geomat Inform Sci Wuhan Univ. 2020;45(5):753–9.
2. Liu X, Ren T, Sun X, Xu C, Zhou M. A review of three dimensional digitalization methods for Ancient Chinese Architecture. J Shanxi Univ (Nat Sci Ed). 2023;3:592–603.
3. Zhao J, Hua X, Yang J, Yin L, Liu Z, Wang X. A review of point cloud segmentation of architectural cultural heritage. ISPRS Ann Photogramm Remote Sens Sp Inform Sci. 2023;10:247–54.
4. Hu Q, Yang B, Xie L, Rosa S, Guo Y, Wang Z, Markham A. RandLA-Net: Efficient semantic segmentation of large-scale point clouds. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR). 2020; 11108–11117.
5. Chen S, Niu S, Lan T, Liu B. Large-scale 3D point cloud representations via graph inception networks with applications to autonomous driving. 2019. arXiv preprint arXiv:1906.11359.