SOLAR CELL SURFACE DEFECT DETECTION BASED ON OPTIMIZED YOLOV5

Abstract

Solar cell surface defect detection plays a crucial role in ensuring the efficiency, reliability, and quality of photovoltaic (PV) systems. Traditional manual inspection methods are time-consuming, labor-intensive, and often unable to detect microscopic defects accurately. To address these limitations, this study proposes an optimized YOLOv5-based deep learning framework for automatic solar cell surface defect detection. The proposed system utilizes advanced image processing and object detection techniques to identify various surface defects such as cracks, scratches, broken grids, black spots, and contamination in solar cells. YOLOv5 is optimized using enhanced feature extraction, data augmentation, hyperparameter tuning, and lightweight network modifications to improve detection accuracy and real-time performance. The model is trained on high-resolution electroluminescence and infrared solar panel images to achieve robust defect classification under different environmental and lighting conditions. Experimental results demonstrate that the optimized YOLOv5 model provides higher precision, faster detection speed, and improved reliability compared to traditional machine learning and standard object detection methods. The proposed approach helps reduce manufacturing errors, improve solar panel quality control, minimize maintenance costs, and increase the overall efficiency and lifespan of photovoltaic systems