Datasets Comparison

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 SSRN: https://dx.doi.org/10.2139/ssrn.5079228 We encourage expanding our dataset by adding more plants growth data other than Batavia lettuce, for more collaboration please contact omar.o.shalash@aast.edu or prof.mail.metwalli@gmail.com or n.abass@plugngrow.me.

Arab Academy for Science Technology and Maritime Transport

Image Processing, Image Acquisition, Hydroponics

Creative Commons Attribution 4.0 International

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 omar.o.shalash@aast.edu or prof.mail.metwalli@gmail.com or n.abass@plugngrow.me.

Arab Academy for Science Technology and Maritime Transport

Image Processing, Image Acquisition, Hydroponics

Creative Commons Attribution 4.0 International