Accelerating Innovation in Healthcare through High-Performance Computing: Applications, Challenges and Future Perspectives
Accurate estimation of wheat yield is essential for ensuring food security, especially given wheat’s role in providing around 20 % of global calories and protein. Traditional yield estimation often relies on manual counting of wheat ears, a method that is labour-intensive, time-consuming, and impractical for large-scale production. To address these limitations, modern agriculture is increasingly turning to advanced technologies such as remote sensing, drones, and machine learning, which enable more efficient and precise monitoring of crop growth and yield potential.In this context, the present study introduces an automated ear-counting approach that applies machine learning to high-resolution images captured by unmanned aerial vehicles (UAVs). Drone imagery was collected during the late growth stage from 15 wheat fields in Bosnia and Herzegovina and processed at a resolution of 1024 × 1024 pixels. Images were manually annotated to mark regions containing wheat ears, resulting in a curated dataset of 556 high-resolution images. For detection, state-of-the-art models including Faster R-CNN, YOLOv8, and RT-DETR were used. While lower-quality images slightly reduced detection accuracy, overall model performance remained strong. This research demonstrates the value of combining UAV-based imaging with machine learning to modernise agricultural practices, offering an efficient, scalable solution for yield prediction and supporting greater sustainability and competitiveness in wheat production.