《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (5): 1694-1702.DOI: 10.11772/j.issn.1001-9081.2024050632

• 多媒体计算与计算机仿真 • 上一篇    

基于单目视觉输电线路精细化巡检方法

王文帅, 韩军(), 胡广怡, 陈炣燏   

  1. 上海大学 通信与信息工程学院,上海 200444
  • 收稿日期:2024-05-17 修回日期:2024-08-20 接受日期:2024-08-22 发布日期:2024-08-29 出版日期:2025-05-10
  • 通讯作者: 韩军
  • 作者简介:王文帅(1996—),男,河南商丘人,硕士研究生,CCF会员,主要研究方向:无人机巡检、深度估计
    韩军(1965—),男,河南三门峡人,副教授,博士,主要研究方向:无人机巡检、计算机视觉
    胡广怡(2001—),女,湖北黄冈人,硕士研究生,主要研究方向:深度估计
    陈炣燏(2000—),女,广西南宁人,硕士研究生,主要研究方向:无人机视觉定位。
  • 基金资助:
    国家自然科学基金资助项目(62371278)

Refined inspection method for power transmission lines based on monocular vision

Wenshuai WANG, Jun HAN(), Guangyi HU, Keyu CHEN   

  1. School of Communication and Information Engineering,Shanghai University,Shanghai 200444,China
  • Received:2024-05-17 Revised:2024-08-20 Accepted:2024-08-22 Online:2024-08-29 Published:2025-05-10
  • Contact: Jun HAN
  • About author:WANG Wenshuai, born in 1996, M. S. candidate. His research interests include UAV inspection, depth estimation.
    HAN Jun, born in 1965, Ph. D., associate professor. His research interests include UAV inspection, computer vision.
    HU Guangyi, born in 2001, M. S. candidate. Her research interests include depth estimation.
    CHEN Keyu, born in 2000, M. S. candidate. Her research interests include UAV visual positioning.
  • Supported by:
    National Natural Science Foundation of China(62371278)

摘要:

针对当前输电线路等空中人造目标的无人机(UAV)精细化巡检轨迹生成方法繁琐、精度不高以及未能以最佳角度拍摄人造目标局部细节等问题,提出一种可以用于输电线路的UAV精细化巡检的实时深度感知与实时线路部件分割定位算法,并构建输电线路单目视觉感知定位导航的最优巡检点路径。通过实时量化调整巡检过程中UAV位置与云台相机拍摄角度,该方法既保证UAV巡检时始终保持安全巡检距离,又使得云台相机能够清晰准确地拍摄包含待巡检目标的图像。采用大疆UAV采集的真实输电线路图像数据和Unreal Engine 4(虚幻引擎)场景下的输电线路图像数据进行实验仿真验证。结果表明,优化的深度感知算法与线路部件分割定位算法能够满足实时性要求。在深度感知与分割定位输出信息的指导下,这些算法能够将UAV位置和云台相机姿态调整为最佳,进而获得高质量的输电线路UAV巡检图像,且最终生成的输电线路精细化巡检轨迹能显著提高运维人员的巡检效率。

关键词: 输电线路部件, 深度感知, 分割定位, 精细化巡检轨迹, 人造目标

Abstract:

Aiming at the current challenges of the complexity, low accuracy, and inability to capture detailed local features of artificial targets from optimal angles in generating refined inspection trajectories for Unmanned Aerial Vehicles (UAVs) inspecting aerial artificial targets such as power transmission lines, a real-time depth perception and line component segmentation and localization algorithm for refined UAV inspection of power transmission lines was proposed, and an optimal inspection point path for monocular vision perception, positioning, and navigation of power transmission lines was constructed. In the method, by adjusting the UAV position and gimbal camera shooting angle quantitatively during the inspection process in real time, a safe inspection distance was maintained while allowing the gimbal camera to shoot images containing the targets to be inspected clearly and accurately. Experimental simulations were carried out by using real data collected by DJI UAV and the data under Unreal Engine 4 scenario. The results demonstrate that the optimized depth perception algorithm as well as the line component segmentation and localization algorithm meets real-time requirements. Under the guidance of the output information from depth perception as well as segmentation and localization, these algorithms can adjust the UAV position and gimbal camera posture optimally, resulting in high-quality UAV inspection images of power transmission lines, and the finally generated refined inspection trajectories can improve the efficiency of inspections of operation and maintenance personnel significantly.

Key words: power transmission line component, depth perception, segmentation and localization, refined inspection trajectory, artificial target

中图分类号: