3D reconstruction of tunnels based on UAV and computer vision techniques

doi:10.11996/JG.j.2095-302X.2026040882

Abstract

Abstract:

Tunnel 3D reconstruction technology can be widely applied in structural defect detection, operation and maintenance state visualization, structural safety assessment, and digital twin model construction. To address the shortcomings of existing monocular vision-based tunnel 3D reconstruction technologies, such as low automation, insufficient modeling accuracy, and poor efficiency, a tunnel 3D reconstruction technique integrating drones and computer vision was proposed. On the hardware end, a tunnel inspection drone was developed, featuring a body structure made of carbon fiber and aluminum alloy, equipped with optical flow sensors, LiDAR, a three-axis gyroscope, a three-axis accelerometer, an electronic compass, and a digital barometer. With a multi-sensor fusion strategy, centimeter-level positioning and navigation were achieved in GNSS-restricted tunnel environments. The drone was also equipped with a high-performance electro-optical pod to enable real-time high-definition video acquisition and transmission. On the algorithm end, a tunnel 3D reconstruction algorithmic framework was designed, comprising video segmentation and frame extraction, camera pose and sparse point-cloud estimation using SfM, depth-map generation via the Gipuma algorithm, surface-model reconstruction based on TGV optimization, and tunnel-texture model construction using texture reconstruction techniques. Experimental results demonstrated that the tunnel inspection drone could stably operate in complex tunnel environments, capturing high-quality video data. Using video frames as the data source, the reconstructed highway tunnel 3D model achieved a root mean square reprojection error of 1.52 pixels, a maximum reprojection error of 25.44 pixels, and a total of 668 088 mesh faces. Further analysis of the impact of video-frame count and resolution on 3D reconstruction revealed that using 99 video frames as the input data achieved a balance between modeling accuracy and efficiency. When computational resources were abundant, it was recommended to upsample the original video frames before performing SfM and then use the original-sized depth maps for surface-model reconstruction. Additionally, a comparison between the proposed modeling framework and traditional solutions showed that the proposed methods exhibit higher computational efficiency and precision, as well as lower dependency on hardware resources. Furthermore, generalization tests in metro tunnels verify its engineering scalability. The proposed equipment and algorithms provide robust support for automated, accurate, and efficient tunnel 3D reconstruction.

Key words: tunnel drone, 3D reconstruction, structure from motion, depth map, implicit modeling

CLC Number:

U456.3

FENG Yong, ZHANG Xiaolei, LV Ruixin, ZHU Shulin. 3D reconstruction of tunnels based on UAV and computer vision techniques[J]. Journal of Graphics, 2026, 47(4): 882-894.

Figures/Tables 23

References 22

[1]	洪开荣, 冯欢欢. 中国公路隧道近10年的发展趋势与思考[J]. 中国公路学报, 2020, 33(12): 62-76. DOI
	HONG K R, FENG H H. Development trends and views of highway tunnels in China over the past decade[J]. China Journal of Highway and Transport, 2020, 33(12): 62-76 (in Chinese).
[2]	陈湘生, 徐志豪, 包小华, 等. 中国隧道建设面临的若干挑战与技术突破[J]. 中国公路学报, 2020, 33(12): 1-14. DOI
	CHEN X S, XU Z H, BAO X H, et al. Challenges and technological breakthroughs in tunnel construction in China[J]. China Journal of Highway and Transport, 2020, 33(12): 1-14 (in Chinese).
[3]	《中国公路学报》编辑部. 中国交通隧道工程学术研究综述·2022[J]. 中国公路学报, 2022, 35(4): 1-40. DOI
	Editorial Department of China Journal of Highway and Transport. Review on China’s traffic tunnel engineering research: 2022[J]. China Journal of Highway and Transport, 2022, 35(4): 1-40 (in Chinese).
[4]	韩兴博, 叶飞, 王永东, 等. 劣化可靠度下的隧道衬砌结构预防性养护决策[J]. 中国公路学报, 2021, 34(1): 104-115. DOI
	HAN X B, YE F, WANG Y D, et al. Predictive maintenance decision-making of tunnel lining structure based on deterioration reliability[J]. China Journal of Highway and Transport, 2021, 34(1): 104-115 (in Chinese).
[5]	FENG Y, FENG S J, ZHANG X L, et al. A two-step deep learning-based framework for metro tunnel lining defect recognition[J]. Tunnelling and Underground Space Technology, 2024, 150: 105832. DOI URL
[6]	FENG S J, FENG Y, ZHANG X L, et al. Deep learning with visual explanations for leakage defect segmentation of metro shield tunnel[J]. Tunnelling and Underground Space Technology, 2023, 136: 105107. DOI URL
[7]	FENG Y, ZHANG X L, FENG S J, et al. Improved SOLOv2 detection method for shield tunnel lining water leakages[J]. Journal of Intelligent Construction, 2023, 1(1): 1-15. DOI URL
[8]	FENG Y, ZHANG X L, FENG S J, et al. Intelligent segmentation and quantification of tunnel lining cracks via computer vision[J]. Structural Health Monitoring, 2025, 24(3) 1896-1926. DOI URL
[9]	MA D, FANG H Y, WANG N N, et al. A low-cost 3D reconstruction and measurement system based on structure-from-motion (SFM) and multi-view stereo (MVS) for sewer pipelines[J]. Tunnelling and Underground Space Technology, 2023, 141: 105345. DOI URL
[10]	徐青山. 基于块匹配和深度学习的多视图立体视觉方法研究[D]. 武汉: 华中科技大学, 2021.
	XU Q S. Research on PatchMatch and deep learning for multi-view stereo[D]. Wuhan: Huazhong University of Science and Technology, 2021 (in Chinese).
[11]	李子欣. 基于Metashape的移动机器人三维地面环境的重构[D]. 北京: 北京交通大学, 2019.
	LI Z X. Reconstruction of 3D navigation environment for a mobile robot based on Metashape[D]. Beijing: Beijing Jiaotong University, 2019 (in Chinese).
[12]	葛海明, 张维, 王小龙, 等. 基于SfM的城市电缆隧道三维重建方法优化研究[J]. 图学学报, 2023, 44(3): 540-550. DOI
	GE H M, ZHANG W, WANG X L, et al. Optimization of 3D reconstruction method for urban cable tunnel based on SfM method[J]. Journal of Graphics, 2023, 44(3): 540-550 (in Chinese).
[13]	SCHÖNBERGER J L, FRAHM J M. Structure-from-motion revisited[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2016: 4104-4113.
[14]	史沛哲. 基于三维重建的盾构隧道病害检测与评估方法研究[D]. 上海: 同济大学, 2022.
	SHI P Z. Research on defect inspection and evaluation method of shield tunnel based on 3D reconstruction[D]. Shanghai: Tongji University, 2022 (in Chinese).
[15]	XUE Y D, ZHANG S, ZHOU M L, et al. Novel SfM-DLT method for metro tunnel 3D reconstruction and visualization[J]. Underground Space, 2021, 6(2): 134-141. DOI URL
[16]	XUE Y D, SHI P Z, JIA F, et al. 3D reconstruction and automatic leakage defect quantification of metro tunnel based on SfM-Deep learning method[J]. Underground Space, 2022, 7(3): 311-323. DOI URL
[17]	FANG K, DONG A, TANG H M, et al. 3D tunnel reconstruction and visualization through multi-smartphone photogrammetry[J]. Measurement, 2023, 223: 113764. DOI URL
[18]	GALLIANI S, LASINGER K, SCHINDLER K. Massively parallel multiview stereopsis by surface normal diffusion[C]// 2015 IEEE International Conference on Computer Vision. New York: IEEE Press, 2015: 873-881.
[19]	POLIARNYI N. Out-of-core surface reconstruction via global TGV minimization[C]// 2021 IEEE/CVF International Conference on Computer Vision. New York: IEEE Press, 2021: 5621-5630.
[20]	LU Y J, WANG S, FAN S S, et al. Image-based 3D reconstruction for multi-scale civil and infrastructure projects: a review from 2012 to 2022 with new perspective from deep learning methods[J]. Advanced Engineering Informatics, 2024, 59: 102268. DOI URL
[21]	LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110. DOI
[22]	FISCHLER M A, BOLLES R C. Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography[J]. Communications of the ACM, 1981, 24(6): 381-395. DOI URL

无人机参数	单位	取值
尺寸	mm	≤990(含四向可碰撞保护罩)
无刷电机数量	个	4
空机重量	g	≥2 800(含四向可碰撞保护罩)
最大起飞重量	g	≥4 900(含电池及保护罩)
最大飞行时间	min	≥15
最大飞行速度	m/s	2.0
最大上升速度	m/s	1.0
工作温度	℃	-20~50
最大操控距离	m	≥300(无遮挡)

无人机参数	单位	取值
尺寸	mm	≤990(含四向可碰撞保护罩)
无刷电机数量	个	4
空机重量	g	≥2 800(含四向可碰撞保护罩)
最大起飞重量	g	≥4 900(含电池及保护罩)
最大飞行时间	min	≥15
最大飞行速度	m/s	2.0
最大上升速度	m/s	1.0
工作温度	℃	-20~50
最大操控距离	m	≥300(无遮挡)

激光雷达参数	单位	描述
激光波长	nm	905
量程	m	40(目标物体反射率10%) 70(目标物体反射率80%)
近处盲区	m	0.1
视场角	°	360(水平) -7~52(竖直)
测距随机误差	cm	≤2(@10 m) ≤3(@0.2 m)
角度随机误差	°	≤0.15
点云输出	点/秒	200 000
尺寸	mm	65×65×60
重量	g	265

激光雷达参数	单位	描述
激光波长	nm	905
量程	m	40(目标物体反射率10%) 70(目标物体反射率80%)
近处盲区	m	0.1
视场角	°	360(水平) -7~52(竖直)
测距随机误差	cm	≤2(@10 m) ≤3(@0.2 m)
角度随机误差	°	≤0.15
点云输出	点/秒	200 000
尺寸	mm	65×65×60
重量	g	265

光电吊舱参数		单位	描述
云台参数	角度抖动量	°	±0.01
	可控俯仰转动范围	°	-135~+45
	可控水平转动范围	°	-160~+160
	横滚转动范围	°	-30~+30
相机参数	镜头焦距	mm	5.15~47.38(±5%)
	光圈	F	1.8~2.5
	视场	°	无变倍：对角79.5，水平71.5 10×光学变倍：对角7.7，水平6.7
	TF卡录视频分辨率	—	2 K (2 560×1 440)@30 fps 1 080 p(1 920×1 080)@30 fps 720 p(1 280×720)@30 fps
	视频存储码率	Mbps	12
	视频文件格式	—	MP4