Assessing the contamination intensity of Porcelain insulators using deep learning networks by UAV

Environmental factors such as dust storms and pollution can lead to corrosion and aging of equipment, which can cause power outages. Contamination flashovers, in particular, have become a major threat to power grid reliability. To prevent such events, it is necessary to accurately assess the contamination intensity of insulators. Unmanned Aerial Vehicles (UAVs) provide a practical means of collecting data for this purpose. In this study, a deep learning model is used to analyze images captured by a drone camera to classify the level of contamination on insulators using the ESDD (Equivalent Density of Salt Deposition) criterion. The dataset is augmented using the CYCLE GAN method to increase its size. Two object detection algorithms are trained and evaluated for accuracy and speed on the same hardware, with YOLOv5 selected as the best trade-off between the two factors. After fine-tuning on a local dataset, the model achieved accuracy above 90% on test data. This approach has the potential to improve the reliability of power systems and minimize the impact of contamination flashovers.