基于融合拉普拉斯残差图像处理的输电线路巡检单目测距技术

doi:10.3969/j.issn.1008-0198.2025.03.017

湖南电力 ›› 2025, Vol. 45 ›› Issue (3): 122-128.doi: 10.3969/j.issn.1008-0198.2025.03.017

• 电力人工智能与数字化 • 上一篇下一篇

基于融合拉普拉斯残差图像处理的输电线路巡检单目测距技术

李勃铖^1,2, 杨昊彦^1,2, 刘钢^1,2, 郭昊^1,2, 刘兰兰^1,2, 李赛³, 席燕辉³

1.国网湖南省电力有限公司超高压输电公司,湖南长沙 410004;
2.带电作业及智能巡检技术国家电网公司实验室（培育）,湖南长沙 410004;
3.长沙理工大学电气与信息工程学院,湖南长沙 410114

收稿日期:2025-03-10 修回日期:2025-04-15 发布日期:2025-07-02
作者简介:李勃铖（1996）,男,工程师,主要从事输变电工程电磁环境、输电线路智能巡检方面的研究工作。杨昊彦（1995）,女,工程师,主要从输电线路智能巡检方面的研究工作。席燕辉（1979）,女,教授,主要从事深度学习方面的研究工作。
基金资助:
国家自然科学基金项目(52277078); 国网湖南省电力有限公司科技项目（5216AJ22000P）; 湖南省自然科学基金项目(2022JJ30609,2021JJ30186)

Transmission Line Inspection Monocular Ranging Technology Based on Fused Laplace Residual Image Processing

Li Bocheng^1,2, YANG Haoyan¹, LiU Gang^1,2, GUO Hao^1,2, LiU Lanlan^1,2, Li Sai³, Xi Yanhui³

1. State Grid Hunan Extra High Voltage Transmission Company, Changsha 410004, China;
2. Live Working and intelligent inspection Technology of State Grid Corporation of China Laboratory (incubation), Changsha 410004, China;
3. School of Electrical and information Engineering, Changsha University of Science and Technology,Changsha 410114, China

Received:2025-03-10 Revised:2025-04-15 Published:2025-07-02

摘要/Abstract

摘要： 在对超、特高压输电线路开展无人机巡检时,无人机能靠近输电线路,但距离过近会增加飞行干扰甚至发生碰撞,因此准确获取无人机与输电线之间的距离至关重要。针对这一问题,提出融合拉普拉斯残差的深度估计网络LapNet（Laplace net）,以U-Net作为网络的基础架构,构建多尺度特征金字塔,提出LUB（Laplace U-Block）模块,替换U-Net的普通卷积层;在LUB模块中引入拉普拉斯金字塔残差结构,增加网络深度的同时保留深度特征中的高频信息,从而引导重建高质量的深度图像,实现单目距离的精确测量。实际测试结果表明,利用无人机单目镜头开展距离检测的平均误差为4.786%,最大误差为8.36%,能够实现无人机镜头与输电线路之间距离的精确测量。

关键词: 单目视觉, 深度估计, 拉普拉斯金字塔, LapNet, 无人机巡检

Abstract: During UAV inspections on ultra-and extra-high-voltage transmission lines, the UAV can approach the lines to capture detailed information. However, too close distance between UAV and transmission lines increases the risk of electromagnetic interference or even collisions. Therefore, it is crucial to accurately obtain the distance between the UAV and the transmission lines. In order to solve this problem, a depth estimation network LapNet(Laplace Net) fused with Laplace residuals is proposed, and U-Net is used as the network infrastructure to construct a multi-scale feature pyramid. A LUB (Laplace U-Block) module is proposed to replace the ordinary convolutional layer of U-Net, and the Laplace pyramid residual structure is introduced into the LUB module to increase the network depth while retaining the high-frequency information in the depth features. This leads to the reconstruction of high-quality depth images for accurate measurement of monocular distances. According to the actual test, the average error of distance detection using UAV monocular lens is 4.786%, and the maximum error is 8.36%, which can realize the accurate measurement of the distance between UAV lens and line.

Key words: monocular vision, depth estimation, pyramid of Laplace, Laplace Net, UAV inspection

中图分类号:

TM755

李勃铖, 杨昊彦, 刘钢, 郭昊, 刘兰兰, 李赛, 席燕辉. 基于融合拉普拉斯残差图像处理的输电线路巡检单目测距技术[J]. 湖南电力, 2025, 45(3): 122-128.

Li Bocheng, YANG Haoyan, LiU Gang, GUO Hao, LiU Lanlan, Li Sai, Xi Yanhui. Transmission Line Inspection Monocular Ranging Technology Based on Fused Laplace Residual Image Processing[J]. Hunan Electric Power, 2025, 45(3): 122-128.

参考文献

[1] EIGEN D,PUHRSCH C,FERGUS R. Depth map prediction from a single image using a multi-scale deep network[EB/OL]. (2014-06-09)[2025-03-19]. https://arxiv.org/abs/1406.2283v1.
[2] 邹德华,张宏伟,江维,等. 基于改进YOLO算法的输电线小目标金具轻量化视觉识别方法[J]. 湖南电力,2025,45(1):136-143.
[3] 杨帆,王梦珺,谭天,等. 基于目标像素宽度识别的电力设备红外成像单目测距改进算法[J]. 电工技术学报,2023,38(8):2244-2254.
[4] 高宏伟,王慧科,刘传银,等. 单相机变焦图像深度估计技术研究[J]. 计算机科学,2014(S1):164-166.
[5] GEIGER A,LENZ P,STILLER C,et al.Vision meets robotics:the KITTI dataset[J]. The International Journal of Robotics Research,2013,32(11):1231-1237.
[6] SILBERMAN N,HOIEM D,KOHLI P,et al.Indoor segmentation and support inference from RGBD images[C]//Computer Vision-ECCV 2012. Berlin,Heidelberg. Springer Berlin Heidelberg,2012:746-760.
[7] KUZNIETSOV Y,STUCKLER J,LEIBE B.Semi-supervised deep learning for monocular depth map prediction[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Honolulu,HI. IEEE,2017:2215-2223.
[8] 张玉亮,赵智龙,付炜平,等. 融合边缘语义信息的单目深度估计[J]. 科学技术与工程,2022,22(7):2761-2769.
[9] SONG M,LIM S,KIM W.Monocular depth estimation using Laplacian pyramid-based depth residuals[J]. IEEE Transactions on Circuits and Systems for Video Technology,2021,31(11):4381-4393.
[10] 蒲正东,陈姝,邹北骥,等. 基于高分辨率网络的自监督单目深度估计方法[J]. 计算机辅助设计与图形学学报,2023,35(1):118-127.
[11] 王泉德,王奇坤,程凯,等. 强化边缘的单目图像深度估计[J]. 华中科技大学学报(自然科学版),2022,50(3):36-42.
[12] LEE J H,HAN M K,KO D W,et al. From big to small:Multi-scale local planar guidance for monocular depth estimation[EB/OL]. (2019-07-24)[2025-03-19]. https://arxiv.org/abs/1907.10326v6.
[13] RONNEBERGER O,FISCHER P,BROX T.U-Net:Convolutional networks for biomedical image segmentation[C]//Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015:18th International Conference. Munich,Germany.Springer:2015:234-241.
[14] FU H,GONG M,WANG C,et al.Deep ordinal regression network for monocular depth estimation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City,UT. IEEE,2018:2002-2011.
[15] MING Y,MENG X Y,FAN C X,et al.Deep learning for monocular depth estimation:a review[J]. Neurocomputing,2021,438:14-33.
[16] GODARD C,MAC AODHA O,FIRMAN M,et al.Digging into self-supervised monocular depth estimation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision(ICCV). Seoul,Korea. IEEE,2019:3827-3837.
[17] GAN Y,XU X,SUN W,et al.Monocular depth estimation with affinity,vertical pooling,and label enhancement[C]//Proceedings of the European Conference on Computer Vision(ECCV). Cham: Springer International Publishing,2018:232-247.
[18] XU H,LI F.Multilevel pyramid network for monocular depth estimation based on feature refinement and adaptive fusion[J]. Electronics,2022,11(16):2615.
[19] WU J P,JI R R,WANG Q,et al.Fast monocular depth estimation via side prediction aggregation with continuous spatial refinement[J]. IEEE Transactions on Multimedia,2023,25:1204-1216.
[20] MANIMARAN G,SWAMINATHAN J.Focal-WNet:An architecture unifying convolution and attention for depth estimation[C]//Proceedings of 2022 IEEE International Conference for Convergence in Technology. Mumbai,India. IEEE,2022:1-7.

基于融合拉普拉斯残差图像处理的输电线路巡检单目测距技术

Transmission Line Inspection Monocular Ranging Technology Based on Fused Laplace Residual Image Processing

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价

[1]	杨洪朝, 席燕辉, 项胜, 徐志康. 基于深度学习的输电线路巡检无人机图像分类方法[J]. 湖南电力, 2024, 44(1): 53-58.
[2]	张家盛, 梁进兴. 基于深度学习的无人机巡检架空输电线路金具锈蚀缺陷检测方法[J]. 湖南电力, 2022, 42(5): 75-78.