隧道巡检机器人自主定位及巡检策略研究  

Research on autonomous positioning and inspection strategy of tunnel inspection robot

作　　者：李程远 郭晓栋[1,2] 常宏 郭世杰 唐术锋[1,2] 张学炜[1,2] LI Chengyuan;GUO Xiaodong;CHANG Hong;GUO Shijie;TANG Shufeng;ZHANG Xuewei(School of Mechanical Engineering,Inner Mongolia University of Technology,Hohhot 010051,China;Inner Mongolia Key Laboratory of Special Service Intelligent Robotics,Inner Mongolia University of Technology,Hohhot 010051,China)

机构地区：[1]内蒙古工业大学机械工程学院,内蒙古呼和浩特010051 [2]内蒙古工业大学内蒙古自治区特殊服役智能机器人重点实验室,内蒙古呼和浩特010051

出　　处：《内蒙古工业大学学报(自然科学版)》2025年第1期17-22,共6页Journal of Inner Mongolia University of Technology:Natural Science Edition

基　　金：国家自然科学基金项目(61763036);内蒙古自治区自然科学基金项目(2021MS05005);内蒙古自治区高等学校创新团队发展计划支持(NMGIRT2213)。

摘　　要：在隧道电缆巡检领域,巡检机器人的运用逐渐增多。针对隧道内信号质量差、巡检机器人自主定位能力弱的问题,提出了一种基于深度相机的机器人巡检自主定位算法。鉴于电缆出现故障时,会产生温度异常这一特点,通过红外热像仪采集温度信息图像,对图像进行预处理,使用深度学习网络对标注好的图像数据集进行训练,生成检测模型。将模型检测图像中温度异常区域作为特征提取异常点,自主定位算法使用这些特征异常点进行定位,实现对隧道内电缆温度异常点精准定位。In the field of tunnel cable inspections,the use of inspection robots is becoming more prevalent.To address issues such as poor signal quality and weak autonomous positioning capabilities within tunnels,inspection robot's autonomous positioning algorithm based on depth cameras is proposed.Given the characteristic of temperature anomalies occurring when cable faults occur,infrared thermal imagers are employed to capture temperature information images.After preprocessing the images,a deep learning network is trained on a labeled data set of images to generate a detection model.The model identifies temperature anomaly regions in the inspected images as feature points,utilizing these points for autonomous positioning.The autonomous positioning algorithm uses these feature points to precisely locate temperature anomalies in the tunnel cables,thereby achieving accurate localization.

关 键 词：隧道巡检 定位 ORB-SLAM2 YOLOv5s 红外热成像 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]