Detection of Box-type Transformer Components Based on Infrared Image Processing
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.
万方数据
维普
