Global date palm production is steadily increasing and adopting technologies such as unmanned aerial vehicles (UAVs) and deep learning can reduce costs, save time, and improve productivity. To address this issue, the authors have proposed an innovative approach that uses UAVs for high-resolution aerial imaging. These images, collected by the Department of Computer Engineering at Al-Salam University in Baghdad and the Institute of Machine Design, Faculty of Mechanical Engineering, Poznan University of Technology, support improved orchard management, palm counting, and yield estimation. Precise spraying and pollination are also facilitated and accelerated, reducing overall cultivation costs. The proposed methodology involves processing captured images and applying three versions of the You Only Look Once (YOLO) object detection algorithm, v11, v12, and YOLO-NAS—to determine the most effective model. The YOLOv12 model achieved the highest mAP@50 at 99.12%, which validates its superior performance in this application. The main innovation is the integration of deep learning-based palm crown detection with UAV imagery, enabling automated and scalable monitoring of palm plantations. The proposed methodology enables rapid, cost-effective, and scalable palm tree enumeration and management. A mobile application based on the trained model is planned to support real-time palm detection, yield estimation, and resource optimisation for farmers and stakeholders.
