Novel deep neural networks for inspection of solar panel anomalies using multimodal images/videos

To optimize Armenia's energy ecosystems, this project proposes cost-effective multimodal systems for solar panel analysis. Manual data processing is inefficient, and computerized monitoring is challenged by poor illumination and changing weather. A novel multimodal AI system will be developed to enhance efficiency and productivity in managing solar panels. The project will evaluate advanced deep learning architectures to detect, classify, and prioritize anomalies using still and thermal images. It will utilize infrared benchmarking datasets and a new dataset capturing anomalies specific to Armenia's climate, dust, and wind conditions.