HydroGrowNet of Batavia Dataset

Over a three-month period, we conducted three consecutive 30-day experiments to monitor Batavia lettuce growth under varying conditions (water temperature, electrical conductivity, and pH). Each day, a dual-camera system traversed the hydroponic channels, capturing high-resolution images of every plant. In total, the dataset contains over 390,000 images, each annotated with environmental and water-quality metadata, forming a robust resource for machine learning–based growth prediction and anomaly detection in hydroponic farming. This data is provided after segmentation of yolo-v8.

The HydroGrowNet dataset is a comprehensive multi-modal dataset designed for hydroponic farming research, particularly for plant growth prediction, anomaly detection, and environmental impact analysis. It integrates sensor-based environmental and water quality measurements with high-resolution plant growth images, providing a rich dataset for machine learning (ML) applications in controlled environment agriculture (CEA).

The dataset was collected from a Nutrient Film Technique (NFT) hydroponic system cultivated with Batavia lettuce (Lactuca sativa L.) over a three-month period. It consists of structured numerical sensor data and unstructured image data, synchronized via timestamps, allowing for detailed growth analysis and fusion-based machine learning applications.
Please check and cite our publication at DOI: 10.1016/j.engappai.2025.111214

We encourage expanding our dataset by adding more plants growth data other than Batavia lettuce, for more collaboration please contact [e-mail] or [e-mail] or [e-mail].