Refined 3D Gaussian Splatting Method for Distribution Network Modeling Towards AR-Oriented Operation and Maintenance
DOI:
https://doi.org/10.4108/ew.11921Keywords:
Distribution Network, 3D Reconstruction, Digitization, 3D Gaussian Splatting, Augmented Reality (AR)Abstract
Aiming at the problems of missing slender structure modeling, complex background interference, and algorithm efficiency bottlenecks in the 3D reconstruction of distribution networks, this paper proposes an improved 3D Gaussian splatting-based algorithm for 3D reconstruction of distribution network scenes. First, through a semantics-guided density control strategy combined with adversarial background suppression, power equipment and vegetation artifacts are effectively separated. Second, a direction-sensitive covariance optimization method is designed to enhance the geometric continuity and detail fidelity of ultra-slender structures such as power lines and utility poles. Finally, a lightweight progressive splatting framework is constructed to achieve real-time rendering at ≥35 fps under 1080p resolution. Experiments show that the improved algorithm significantly increases the completeness rates of power lines, utility poles, and insulators to 90%, 95%, and 80%, respectively, reduces background artifacts to 2% in area, and shortens the reconstruction time to 0.4 hours. This study provides a high-precision modeling tool for digital twins of distribution networks and supports real-time applications of augmented reality (AR) technology in inspection navigation, fault localization, and remote collaboration.
Downloads
References
[1] Wu, Y. Research on Strategies for Digital Transformation of Distribution Network[J]. Automation Application, 2024,65(16):72-74.
[2] Liu, Z.; Pan, F.; Li, F. Intelligent power distribution network digital technology and its application[J]. Lamps & Lighting, 2025,(01):207-209.
[3] Zhao, Z.; Li, X.; Tong, C. Application Practice of Digital Twin Technology in Distribution Network for New Power System[J]. Popular Utilization of Electricity,2023,38(09):59-62.
[4] Sheng, K.; Wu, X.; Li, Z. Application Architecture of New Urban Power System Based on Digital Twin Technology[J].Electric Engineering,2023,(06):100 -103.
[5] Zhang, Y.; Feng, Y.; Lan, Y. Monitoring Method for Operating Status of 10kV Distribution Network Power Equipment Based on Digital Twin Technology[J]. Scientific and Technological Innovation,2022,(36):163-166.
[6] Yu, J.; Zhang, H.; Qi, H.; et al. Application of Digital Twin Technology in Intelligent Operation and Inspection Business of Power Grid[J]. Geospatial Information,2025,23(03):109-113.
[7] Chen, S. Fine modeling of distribution network equipment based on close proximity photogrammetry technology[J]. Mechanical and Electrical Information,2022,(14):84-88.
[8] Xu, J.; Gao, H. Design of Digital Twin Base System for Power Equipment in Distribution System[J]. Electric Power Equipment Management,2024,(18):170- 172.
[9] Wei, S. Research on Insulator Detection and Defect Recogniton of Distribution Network Based on Deep Learning[D]. Guangxi University,2023.
[10] Liu T.; Liu, H. Discussion on Wind Deflection Technology for Large Span Bare Conductors in Distribution Networks[J]. Rural Electrification,2023,(12):1-4.
[11] Li, H. Safety Management of Ground Distance for Distribution Network Lines[J]. Yunnan Electric Power,2004, (05):44-46.
[12] Miao, J.; Xia, Y.; Sun, M.; et al. Application of three-dimensional point cloud processing in the field survey of non-stopping operation of distribution network[J]. China High and New Technology,2025,(02):58-59+91.
[13] Hu, Z.; Yuan, J.; Gong, Y.; et al. Reconstruction of Lead Wires of Power Lines for Live-line Working Robots in Distribution Networks[J]. Control Engineering of China,2021,28(11):2123-2130.
[14] Wang, Z.; Xue, F.; Liu, K. Efficient three-dimensional reconstruction method of SAR image based on neural radiation field[J/OL]. Computer Measurement & Control,1-18.
[15] Wang, J.; Li, Z.; Wang, H.; et al. Highly Robust Multi-view Reconstruction Network for Neural Radiation Field[J]. Geospatial Information,2025,23(04):16-19+ 99.
[16] Shen, Y.; Yang, Q.; Gong, X.; et al. Research on 3D structure reconstruction of power equipment based on multi-feature point fusion[J/OL]. Computer Technology and Development,1-8.
[17] Wang, W.; Tang, B.; Gu, Z.; et al. Overview of Multi-View 3D Reconstruction Techniques in Deep Learning[J]. Computer Engineering and Applications,2025, 61(06):22-35.
[18] Dong, X.; Qu, F.; Zhang, J.; et al. Key design for 3D reconstruction of overhead transmission line corridors based on oblique photography[J]. China High-Tech Enterprises,2016,(18):13-14.
[19] Yu, W.; Lu, J.; Cheng, H. Review of NeRF-based SLAM Research[J]. Computer Systems & Applications,2025, 34(04):18-33.
[20] Wu, Y.; Li, F.; Yu, T.; et al. Power Data Compression and High-precision Reconstruction Based on Residual Dual Attention Mechanism Network[J]. Power System Technology,2022,46(08):3257-3271.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Conghan Liu, Lijuan Yan, Linjun Lu, Huazhang Tan
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open-access article distributed under the terms of the Creative Commons Attribution CC BY 4.0 license, which permits unlimited use, distribution, and reproduction in any medium so long as the original work is properly cited.
