Microgreen Master Dataset (Public Version): RGB Image Dataset for Classification of Healthy, Dry, and Mold-Affected Microgreens

Ganna Zavolodko, Volodymyr Andriushchenko

The Microgreen Master Dataset (Public Version) is an open RGB image dataset developed to support research in computer vision, artificial intelligence, and smart agriculture. The dataset is intended for training and evaluating machine learning models for automatic classification of the physiological condition of microgreens cultivated in controlled indoor environments. This dataset was created within the Bachelor's research project "Development of an Intelligent IoT System for Microgreen Cultivation Based on Computer Vision" at the Department of Multimedia and Internet Technologies and Systems, National Technical University "Kharkiv Polytechnic Institute" (Ukraine). The research concept, experimental methodology, and dataset design were developed by Hanna Zavolodko, who supervised the research. Image acquisition, dataset preparation, and dataset organization were carried out by Vitalii Andriushchenko during the implementation of his Bachelor's thesis. The published dataset is a representative public subset of the complete research dataset and has been prepared for open distribution while complying with repository file size limitations. The dataset combines images derived from publicly available datasets with original photographs obtained during controlled indoor microgreen cultivation experiments. The original images extend the diversity of the dataset and improve the representation of real cultivation conditions. The source datasets include: Fungus Detection Dataset (Roboflow Universe): https://universe.roboflow.com/moss-detection/fungus-detection-nqyka Wet vs Dry Plant Dataset (Roboflow Universe): https://universe.roboflow.com/plant-moisture-detection/wet-vs-dry-plant These datasets were supplemented with original RGB images collected by the authors. Dataset Statistics Class Images Healthy 52 Dry 6 Mold 390 Total 448 Image format: JPEG (.jpg) Directory structure: training dataset organized by class labels. Applications The dataset can be used for: image classification; computer vision research; deep learning; transfer learning; plant health monitoring; plant disease detection; smart agriculture; precision agriculture; intelligent IoT systems; educational and research projects. Notes The complete research dataset contains additional images collected during the experimental study. The present release represents a curated subset intended for open scientific dissemination and reproducibility of research results. Acknowledgement If you use this dataset in your research, please cite this Zenodo record. When appropriate, please also acknowledge the original Roboflow Universe datasets used as source material for creating the derived dataset. The dataset incorporates and extends data derived from the following publicly available resources: Fungus Detection Dataset (Roboflow Universe): https://universe.roboflow.com/moss-detection/fungus-detection-nqyka Wet vs Dry Plant Dataset (Roboflow Universe): https://universe.roboflow.com/plant-moisture-detection/wet-vs-dry-plant These publicly available datasets were supplemented with original RGB photographs collected by the authors under controlled laboratory conditions.