摘要
针对当前特高压变电设备缺陷图像识别准确率偏低的问题,文章提出一种基于卷积神经网络(CNN)的新型识别方法。通过构建多层CNN结构,实现了对复杂背景图像中缺陷特征的自动提取与分类。实验结果表明,该方法在多类典型缺陷识别中具有较高的准确率和鲁棒性,能够有效提升检测效率与系统智能化水平。
To address the issue of low accuracy in defect image recognition for ultra-high voltage substation equipment,this article proposes a novel recognition method based on Convolutional Neural Network(CNN).A multi-layer CNN architecture is constructed to achieve automatic extraction and classification of defect features from images with complex backgrounds.Experimental results demonstrate that the proposed method exhibits high accuracy and robustness in recognizing multiple typical defect types,effectively improving detection efficiency and the level of system intelligence.
作者
许守法
武昊天
XU Shoufa;WU Haotian(Ultra-high Voltage Company of State Grid Shandong Electric Power Company,Jinan 250000,China)
关键词
CNN
特高压
变电设备
缺陷识别
CNN
ultra-high voltage
substation equipment
defect recognition
