Nitrogen deficiency in maize: Annotated image classification dataset.

Area of Science:

• Agricultural Science
• Computer Science
• Plant Physiology

Background:

• Nitrogen is crucial for maize yield, but deficiency impairs photosynthesis and reduces crop output.
• Visible symptoms of nitrogen deficiency in maize appear during the growing season.
• Optimizing nitrogen application is key for sustainable agriculture and reducing environmental impact.

Purpose of the Study:

• To develop a dataset for detecting nitrogen deficiency in maize using RGB imagery.
• To enable early detection of nitrogen deficiency for improved crop management.
• To support the optimization of nitrogen fertilizer use in maize production.

Main Methods:

• A dataset of 1200 maize canopy RGB images was created from field trials.
• Trials included three nitrogen fertilization levels: none (N0), 75 kg/ha (N75), and 136 kg/ha (NFull).
• Images were captured using a DSLR camera at a 45° angle, with specific camera settings (f/8, ISO400).

Main Results:

• The dataset contains annotated images of maize under varying nitrogen conditions.
• It includes 238 different maize genotypes across 1200 plots.
• This resource facilitates the development of machine learning models for nitrogen deficiency detection.

Conclusions:

• The developed dataset is valuable for researchers, data scientists, and agronomists.
• It supports advancements in precision agriculture technologies like robotics and UAVs.
• Enables in-season detection of nitrogen deficiency in maize, optimizing fertilizer use and environmental outcomes.