Paper Title
YOLOV5 AND DATA AUGMENTATION: A DUAL APPROACH TO ENHANCING PV ANOMALY DETECTION

Abstract
Solar energy has especially attracted interest because it is cheaper and safer in terms of environmental impacts than any other source of energy. By applying the concepts of advanced monitoring and analytics, any form of anomalies discovered on solar energy systems is quickly detected and resolved, hence enhancing the systems' reliability and efficiency, minimizing overall breakdown, and prolonging effective maintenance intervals. This synergy not only encourages solar energy generation reliability but also plays a role in the process of switching to a cleaner and safer energy future. YOLOv5, one of the most advanced target detection algorithms available, depends heavily on the training dataset for best results. Therefore, its performance will be poor if the number of samples with defects is low. Using data augmentation to raise the quantity of training samples is a practical way to improve YOLOV5 performance. This paper presents a combined YOLOV5 and data augmentation model as a framework to detect the defects on the PV panels. The results show that after implementing the use of data augmentation, the precision, recall and mAP scores of the detection performance of YOLOv5 are 82%, 81.2% and 89.1%, respectively. Keywords - PV panels, YOLO, Data Augmentation, Anomaly Detection.