Solar Panel Damage Detection Using a Drone and YOLO-Supported Artificial Intelligence Model

In recent years, the importance of renewable energy sources has been increasing, and ensuring the efficient operation of large-scale solar farms requires regular maintenance and monitoring. Since traditional manual maintenance methods are time-consuming, costly, and prone to errors, this study aims to detect physical and electrical anomalies in solar panels automatically using drone-based imaging systems and AIsupported analysis methods. The proposed system integrates RGB and thermal imaging techniques to enhance the efficiency of maintenance processes in large-scale solar farms, enabling autonomous inspections over vast areas. The captured images are analyzed using the YOLOv9 object detection algorithm, allowing for the automatic classification of panel conditions. By detecting both physical and electrical faults, the system provides a more efficient maintenance management approach, reduces costs, and minimizes energy losses in solar farms. The performance of the developed system has been evaluated in terms of accuracy, false positive, and false negative rates for different types of deformations, demonstrating significant improvements in largescale solar plant maintenance compared to traditional methods. However, certain limitations were identified in detecting smallscale deformations, and further improvements through training the model on larger datasets have been suggested. The results indicate that drone-based imaging and AI-assisted analysis systems offer a faster, cost-effective, and more reliable alternative for large-scale solar farm maintenance processes.