This paper aims to evaluate the effectiveness of two object detection models, specifically aiming to identify the superior model for detecting photovoltaic (PV) modules based on aerial images. In this study, we examined the deep learning-based YOLOV5n and YOLOV8 models as two prominent YOLO methodologies for PV panel detection. We began by acquiring a dataset comprising 646 images with corresponding binary masks. These images were subsequently partitioned into 506 training images, 140 validation images, and a test set of novel images. For each image, we have created two distinct annotation types: (i) COCO JSON annotations and (ii) YOLO annotations. The models were then trained on our proprietary database, resulting in box precision values of 93.8% for YOLOV5n and 96.7% for YOLOV8.