基于机器视觉的变电站智能巡检机器人的设计与应用

Design and Application of Intelligent Inspection Robots for Substations Based on Machine Vision

针对变电站巡检工作中存在的工作强度大、安全隐患多等问题,设计了一种基于机器视觉的智能巡检机器人系统。首先,该系统引入了一种用于缺陷检测的基于注意力机制的多尺度卷积神经网络,实现了较高的变电站设备缺陷定位和分类精度。同时,设计了一种轻量化的图像增强算法,有效提升恶劣环境下的视觉感知鲁棒性。实验表明,所提机器视觉方法可显著提升变电站巡检的自动化和智能化水平。

In response to the issues of high workload and numerous safety hazards in power substation inspection work,this paper designs an intelligent inspection robot system based on machine vision.Firstly,the system introduces a multi-scale convolutional neural network based on attention mechanisms for defect detection,achieving high precision in the localization and classification of power substation equipment defects.Additionally,this paper designs a lightweight image enhancement algorithm that effectively improves the robustness of visual perception in adverse environments.Experiments demonstrate that the machine vision method can significantly enhance the automation and intelligence level of power substation inspections.