基于红外图像处理的箱式变压器组件检测研究  

Detection of Box-type Transformer Components Based on Infrared Image Processing

作　　者：全欣宇 彭诗援 钱宏杰 Quan Xinyu;Peng Shiyuan;Qian Hongjie(Huaneng New Energy Co.,Ltd.Yunnan Branch,Kunming,China)

机构地区：[1]华能新能源股份有限公司云南分公司,云南昆明

出　　处：《科学技术创新》2024年第14期51-54,共4页Scientific and Technological Innovation

摘　　要：为解决高原风电箱变巡视问题,减少巡视工作量,提高设备运行稳定性,针对基于红外图像处理的箱式变压器组件检测技术展开研究。构建YOLOv8模型作为本文实验的主要算法,在红外图像基础上,通过特征提取、目标检测头部网络、位置预测等步骤,设计检测算法,配置实验参数,训练模型。最后的实验结果表明,YOLOv8n模型mAP达到95.1%,在实验设备上识别速度达到0.35 s,与YOLOv3识别速度相比提升了74%,证明了YOLOv8算法在实际运用中可行性和有效性。To solve the inspection problem of Gaoyuan wind power box transformer,reduce inspection workload,and improve equipment operation stability,research is conducted on the detection technology of box transformer components based on infrared image processing.The YOLOv8 model is constructed as the main algorithm for this article's experiment.Based on infrared images,detection algorithms are designed,experimental parameters are configured,and the model is trained through steps such as feature extraction,target detection head network,and position prediction.The final experimental results show that the mAP of the YOLOv8n model reaches 95.1%,and the recognition speed on the experimental equipment reaches 0.35 s,which is 74%higher than the recognition speed of YOLOv3.This proves the feasibility and effectiveness of the YOLOv8 algorithm in practical application.

关 键 词：箱式变压器部件检测 红外图像识别 深度学习 YOLOv8 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]