Antenna Detection from Satellite Images using Enhanced YOLO

Identifying antenna structures from satellite images is essential for managing telecommunications infrastructure. Traditional object detection methods, including edge detection, template matching, and machine learning models like SVMs, Random Forests, and CNNs, often lack accuracy and efficiency in complex urban environments. This study proposes an enhanced YOLOv10 model, trained on a well-prepared satellite image dataset with preprocessing and augmentation techniques. The model achieved 91.7% accuracy and a false positive rate below 7%, demonstrating effective performance in both urban and rural settings.