Peruvian-Cotton - Annotated Dataset of Gossypium barbadense Fruits in Peruvian Crops

This dataset contains images and annotations used in the study “Automated Detection and Counting of Gossypium barbadense Fruits in Peruvian Crops Using Convolutional Neural Networks.”
When using this dataset, please cite:
Ballena-Ruiz, J.; Arcila-Diaz, J.; Tuesta-Monteza, V. Automated Detection and Counting of Gossypium barbadense Fruits in Peruvian Crops Using Convolutional Neural Networks. AgriEngineering 2025, 7, 152. https://doi.org/10.3390/agriengineering7050152

Data collection was carried out in two cotton crop plots located in the district of Lambayeque, province of Lambayeque, Lambayeque region, Peru (6°40′20″ S, 79°53′17″ W). A total of 370 original images were captured in JPG format at a resolution of 640 × 640 pixels, using a Samsung A12 device mounted on a 2.5 m-high structure.
Out of these, 100 images were selected for manual annotation using LabelImg (v1.8.6), resulting in 3062 YOLO-format annotations corresponding to the class “cotton”.
Subsequently, data augmentation techniques were applied (progressive rotations ranging from −5° to +5°), which expanded the dataset to 2186 images with a total of 70,348 annotated cotton fruits.
This dataset is intended to support research in computer vision, object detection, and agricultural applications based on artificial intelligence.