Solar panels that use photovoltaic technology have a very low carbon footprint. However heavy rain can reduce their power output by over 50% by making the panels dirty. To address this problem, a new system has been developed to measure how much solar power is lost from dirt on the panels. The system uses a model that has been trained to precisely identify clean and dirty sections of solar panels in images. It marks the boundaries between clean and dirty areas and labels each pixel as clean or dirty. This gives detailed information about where and how much dirt is on the panels, including patterns and variations. By using segmentation and classification models, the system gives a comprehensive understanding of solar panel dirt. The segmentation model provides detailed data on the type and location of dirt. The classification model summarizes how much dirt reduces energy generation. The classification model had an output with the pinball testing loss of 0.23.