Dataset for weed detection in fruit orchards

Weed control in permanent crops remains a costly and environmentally demanding task. While UAV-based imaging has advanced rapidly, publicly available datasets for weed detection in orchard settings are scarce, particularly those including multispectral and multi-temporal acquisitions. This lack of resources limits the development of robust machine learning models for cross-spectral adaptation and seasonal generalization. We present CampanetaWeed, a multispectral image dataset acquired with a DJI Mavic 3 Multispectral UAV over a commercial fruit orchard in Corbera, Valencia, Spain. The dataset comprises three flights (October 2023, December 2023, April 2025), providing pixel-aligned RGB and four narrow-band images (R, G, Red-Edge, NIR) at 0.65 cm/pixel resolution (GSD). All images were annotated in YOLOv5 format for six weed species and one ground disturbance class, totaling over 10,000 images and 271,000 labeled objects. CampanetaWeed uniquely combines multispectral alignment, exhaustive annotations, and multi-season coverage, enabling advances in weed detection, temporal robustness, and sustainable precision agriculture.