摘要
针对变电站主设备红外图像的自动识别需求,文中构建了一种面向智能运维的图像识别方法。该方法通过图像预处理提升红外图像质量,结合卷积神经网络实现对设备区域的精准定位与特征提取,引入多尺度特征融合与注意力机制增强识别鲁棒性与准确率。在真实红外图像数据集中进行的实验显示,模型在缺陷识别、目标分类等任务中具有优良性能,可有效支持电力设备智能巡检系统的建设。
For the automatic recognition of infrared images of substation main equipment,an image recognition method for intelligent operation and maintenance is constructed in this paper.The method improves the quality of infrared image by image preprocessing,and realizes the accurate positioning and feature extraction of equipment area by combining convolutional neural network.The multi-scale feature fusion and attention mechanism are introduced to enhance the robustness and accuracy of recognition.Experiments on real infrared image data sets show that the model has excellent performance in defect recognition,target classification and other tasks,and can effectively support the construction of intelligent inspection system for power equipment.
作者
贾伟
JIA Wei(Yan̓an Power Supply Company of State Grid Shaanxi Electric Power Co.,Ltd,Yan̓an,Shaanxi 716000,China)
出处
《移动信息》
2025年第12期89-92,共4页
Mobile Information
关键词
红外图像识别
智能运维
主设备缺陷检测
Infrared image recognition
Intelligent operation and maintenance
Main equipment defect detection
