UAV based Tomato Dataset

This dataset contains UAV-acquired tomato field imagery collected for tomato counting and yield prediction research under real agricultural conditions. The images were captured using an unmanned aerial vehicle (UAV) equipped with high-resolution RGB and/or multispectral sensors over multiple tomato cultivation plots during different crop growth stages, including flowering, fruit setting, ripening, and harvest stages. The dataset includes raw aerial images, orthomosaic patches, and manually annotated tomato instances generated using Roboflow annotation tools. The annotations are provided in object detection format suitable for deep learning applications such as YOLO, COCO, or Pascal VOC. Each image is associated with metadata including flight altitude, overlap settings, environmental conditions, GPS coordinates, and plot-level yield measurements. The primary purpose of the dataset is to support research in: -- Tomato counting using computer vision -- UAV-based precision agriculture -- Yield prediction modeling -- Small-object detection in aerial imagery -- Agricultural AI and deep learning applications The dataset was collected under varying illumination conditions and field environments to improve model robustness and generalization. Ground-truth yield measurements were recorded from corresponding plots to enable supervised learning and benchmarking for yield estimation tasks. This dataset can be used by researchers in: -- Precision agriculture -- Remote sensing -- Computer vision -- Machine learning -- Crop phenotyping The dataset is organized into structured directories containing raw images, annotations, metadata files, and experimental documentation to facilitate reproducibility and reuse in future agricultural AI research.