Mango Orchard UAV Image Dataset (RGB and Multispectral OCN)

This dataset contains UAV-captured image data of mango orchards collected under real field conditions for object detection research. The dataset is organized into two independent subsets based on sensor modality: RGB and multispectral. The RGB subset consists of images captured using an RGB camera mounted on a UAV. The multispectral subset consists of images captured using a multispectral camera providing Orange, Cyan, and Near-Infrared (OCN) bands. The two subsets are not paired or co-registered; they were captured independently but annotated using identical class definitions. Each subset is provided as a self-contained archive that includes images, YOLO-format annotation files, and a data.yaml configuration file. Images are provided in their original captured resolution (e.g., up to 4000 × 3000 pixels) without resizing. Image resizing, preprocessing, and training procedures are described in the associated research article. The dataset is intended to support reproducible research in UAV-based agricultural monitoring and object detection. Users are requested to cite the associated research article and the Zenodo dataset DOI when using this dataset. The research code used for training, validation, and evaluation is publicly available via a separate Zenodo record: https://doi.org/10.5281/zenodo.18516891