Hydroponic Maize Root Image Dataset with Manual Annotations for Semantic Segmentation and Root Phenotyping

This dataset contains original root images of hydroponic maize seedlings and their corresponding manually annotated binary masks. All maize plants were grown in transparent root boxes, and root images were continuously collected using an in situ imaging system. The dataset was constructed to support maize root semantic segmentation, root phenotyping, and related downstream analysis tasks.

The dataset includes data from 11 root boxes, with two maize plants grown in each root box, resulting in a total of 22 hydroponic maize plants. During the experiment, root images were continuously collected over a two-week growth period. Considering that the roots were very small during the early growth stage and provided limited image information, the released dataset mainly includes the original root images and corresponding manually annotated binary masks collected from the 22 hydroponic maize plants during the last 10 days of growth.