Multi-modal fusion-based 3D reconstruction of power corridor using 3DGS

doi:10.13474/j.cnki.11-2246.2026.0321

Abstract

Abstract: Due to the limitations of LiDAR,such as a small field of view and insufficient density of sampling points,the 3D point cloud of power corridors reconstructed on the ground based on LiDAR is prone to information missing.To address this issue,this paper proposes a multi-modal fusion reconstruction method based on Gaussian splatting.Firstly,a SLAM system that fuses vision and LiDAR is adopted to achieve accurate estimation of sensor pose.Secondly,3D Gaussian primitives are initialized,and RGB images and depth maps are rendered through 3D Gaussian splatting (3DGS) technology.Finally,the Gaussian primitives are dynamically updated and the final results are generated by optimizing color and depth errors.Experimental results show that this method can effectively compensate for LiDAR blind areas,complement the reconstruction results using image information,and is superior to traditional methods in both model integrity and accuracy.This paper verifies the effectiveness of 3DGS in multi-modal 3D reconstruction,providing a new and effective approach for 3D reconstruction of power corridors.

Key words: power corridor, 3D reconstruction, multi-modal fusion, 3D Gaussian splatting

CLC Number:

P237

ZHOU Chao, GAO Shuhan, TU Diantao, LIU Yangdong, LIU Hui, SHEN Hao, JIA Ran, LIU Chuanbin, ZHANG Yang, LIU Rong, SHEN Shuhan. Multi-modal fusion-based 3D reconstruction of power corridor using 3DGS[J]. Bulletin of Surveying and Mapping, 2026, 0(3): 124-129.

References

[1] 黄伟.输电线路走廊无人机影像三维重建关键技术研究[D].武汉:武汉大学,2021.
[2] 段敏燕.机载激光雷达点云电力线三维重建方法研究[J].测绘学报,2016,45(12):1495.
[3] 李德友,李彩林,李祥珅,等.激光雷达点云电力线自动提取算法[J].测绘通报,2021(6):33-38.
[4] 陈竹安,邹梓龙,徐志芳,等.复杂背景下航拍图像的电力线自动提取算法[J].测绘通报,2022(4):37-43.
[5] 黄鹤,佟国峰,夏亮,等.SLAM技术及其在测绘领域中的应用[J].测绘通报,2018(3):18-24.
[6] 潘际伟,郑勇峰,汪建林.SLAM激光扫描技术在电力改造工程建筑地下结构探测的辅助应用[J].电力勘测设计,2023(S1):72-75.
[7] KERBL B,KOPANAS G,LEIMKUEHLER T,et al.3D Gaussian splatting for real-time radiance field rendering[J].ACM Transactions on Graphics,2023,42(4):1-14.
[8] ZWICKER M,PFISTER H,VAN BAAR J,et al.EWA splatting[J].IEEE Transactions on Visualization and Computer Graphics,2002,8(3):223-238.
[9] ZHENG Chunran,XU Wei,ZOU Zuhao,et al.FAST-LIVO₂:fast,direct LiDAR-inertial-visual odometry[J].IEEE Transactions on Robotics,2025,41:326-346.
[10] HONG Sheng,HE Junjie,ZHENG Xinhu,et al.LIV-GaussMap:LiDAR-inertial-visual fusion for real-time 3D radiance field map rendering[J].IEEE Robotics and Automation Letters,2024,9(11):9765-9772.
[11] LANG X,LI L,ZHANG H,et al.Gaussian-LIC:Photo-realistic LiDAR-Inertial-Camera SLAM with 3D Gaussian Splatting[C]//Proceedings of 2025 International Conference on Robotics and Automation (ICRA).[S.l.]:IEEE,2025.
[12] MILDENHALL B,SRINIVASAN P P,TANCIK M,et al.NeRF:representing scenes as neural radiance fields for view synthesis[C]//Proceedings of 2020 Computer Vision-ECCV.Cham:Springer,2020:405-421.
[13] BARRON J T,MILDENHALL B,VERBIN D,et al.Mip-NeRF 360:unbounded anti-aliased neural radiance fields[C]//Proceedings of 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New Orleans,LA:IEEE,2022:5460-5469.
[14] LIN Jiarong,ZHANG Fu.R3LIVE:a robust,real-time,RGB-colored,LiDAR-Inertial-Visual tightly-coupled state estimation and mapping package[C]//Proceedings of 2022 International Conference on Robotics and Automation (ICRA).Philadelphia,PA:IEEE,2022:10672-10678.
[15] LIN Jiarong,ZHENG Chunran,XU Wei,et al.R2 LIVE:a robust,real-time,LiDAR-inertial-visual tightly-coupled state estimator and mapping[J].IEEE Robotics and Automation Letters,2021,6(4):7469-7476.