Fusion Between Image Processing and Machine Learning for Dust Detection on Solar Panels

This work explores the impact of environmental factors, especially dust accumulation, on the performance of solar panels in the Arab region, an area particularly prone to dust-related efficiency issues. An innovative approach that combines image processing with advanced machine learning techniques to predict solar panel output under various environmental conditions is introduced. The research involves a detailed examination of factors like dust accumulation, temperature, and solar irradiance, and their collective impact on solar panel efficiency. Two datasets are introduced and a diverse range of machine learning models, including Artificial Neural Networks (ANNs), Long Short-Term Memory networks (LSTMs), and Support Vector Regression (SVR), are evaluated to identify the most effective model for predicting solar panel performance which is proven to be LSTM for the first dataset with an accuracy of 99.50% and ANN for the second dataset with an accuracy of 98.50%. The study's findings are significant for optimizing solar panel maintenance and operation in dust-prone regions, providing valuable insights for enhancing the sustainability and efficiency of solar energy systems in these challenging environments.