TomatoPlantfactoryDataset

The dataset entitled Tomato Plantfactory Dataset, which comprises 520 images of the Micro tomato variety, captured at two different resolutions (6000x4000 and 4032x3024) and at two stages of fruit growth. The images were acquired in a complex lighting environment of an artificial light plant factory, presenting various complexities, including variations in perspectives, lighting quality, distance, and occlusions and blurs of the fruits. The dataset provides a valuable resource for researchers interested in developing new techniques for tomato fruit detection, classification, localization, and visualization. All tomato fruits have been manually annotated, resulting in a total of 9112 instances, including 5996 green fruits and 3116 red fruits. The annotation files are available in both PascalVOC and YOLO formats, facilitating the integration of the dataset into existing computer vision pipelines. Finally, the dataset can be conveniently partitioned for rapid model training purposes.

If you are interested in this dataset and have introduced it in your research, please cite the article of this dataset.
"Z. Wu, M. Liu, C. Sun, X. Wang, A dataset of tomato fruits images for object detection in the complex lighting environment of plant factories, Data in Brief. 48 (2023) 109291. https://doi.org/10.1016/j.dib.2023.109291.
"

A GitHub page for this dataset is also available at:
https://github.com/veveup/Tomato-Plant-Factory-Dataset

The GitHub repository provides a lightweight entry point for the dataset, including a visual sample gallery, dataset description, citation information, related papers using or citing the dataset, and quick-start scripts for checking the local files and preparing YOLO-style train/validation/test splits. The dataset files themselves remain hosted on Mendeley Data as the official source.