“…In addition, the current mainstream YOLOv5 deep learning algorithm can detect about 900 photovoltaic panels in one minute [10] , but this is to remove the time of model training, and ignores the impact of hardware, if the model training time is calculated, YOLOv5 can detect about 400 photovoltaic panel images in one minute, the method proposed in this paper does not need to be trained, and the processing time is much more stable, and it can process in one minute 800 PV panel images. The method proposed in this paper is more based on traditional image processing algorithms, combined with their own needs to improve, compared with the current mainstream machine learning methods, the method proposed in this paper is based on the mathematical model and rules of the data, the algorithm can be interpreted, the computational efficiency of the faster and lower consumption of resources, without the need for a large number of samples to be trained for the identification of minor defects of the photovoltaic panels, this paper's proposed method is also more robust.…”