In order to accurately analyze the ash accumulation on photovoltaic panels

a photovoltaic dust visualization experimental platform was built

and the average grayscale value was introduced to numerically analyze the photovoltaic panel images. The clear correspondence between the average grayscale value of photovoltaic panel images and the dust density of photovoltaic panels was verified. On this basis

five fusion methods were used to fuse the visible light images and infrared images collected from the dual spectral image fusion experimental platform. The five types of fusion images were combined with visible light images and infrared images to form an image dataset. These seven types of images were identified and analyzed. The results showed that

the recognition effect of infrared images on the degree of ash accumulation on photovoltaic panels was the least affected by irradiance

with the highest accuracy

and the most significant change in the degree of ash accumulation was reflected. This conclusion can provide a theoretical basis for the study of ash accumulation rules and is of great significance for the recognition of the degree of ash accumulation on photovoltaic panels.