Improved YOLOv9-Based Insulation Breakage Detection of Column Transformer

In order to safeguard the insulation performance of column-mounted transformers and improve the reliability of distribution network operation, it is proposed to conduct a study on the detection and spraying of insulation breakage of column-mounted transformers based on unmanned aerial vehicles (UAVs).The breakage detection model is selected from YOLOv9 and improved by introducing Haar wavelet downsampling, adding iEMA attention mechanism and replacing MPDIoU loss function to improve the detection accuracy. This model [email protected] Achieved 93.7% and an FPS of 74.3 frames per second, meeting the requirements for insulation damage detection.