Artificial intelligence in plant science: from image-based phenotyping to yield and trait prediction

Area of Science:

• Agricultural Science
• Computer Science
• Remote Sensing

Background:

• Plant research traditionally relies on manual measurements, which are labor-intensive and time-consuming.
• Advancements in artificial intelligence (AI), imaging, and remote sensing offer opportunities for automated data collection.
• High-throughput image-based phenotyping allows for precise, automated trait acquisition at various scales.

Purpose of the Study:

• To present a cross-disciplinary paradigm merging AI with plant phenotyping and yield forecasting.
• To highlight how AI enhances trait monitoring and yield prediction accuracy.
• To promote accurate and sustainable modern agriculture.

Main Methods:

• Utilizing artificial intelligence (AI) for complex imaging and remote sensing data analysis.
• Implementing high-throughput image-based phenotyping for automated trait acquisition.
• Integrating satellite observations, unmanned aerial vehicle (UAV) imaging, soil, and climate data.

Main Results:

• AI enables precise, automated trait acquisition across diverse spatial and temporal scales.
• The integration of multiple data sources (satellite, UAV, soil, climate) improves trait monitoring and yield prediction.
• Enhanced evaluation and prediction of crop performance under variable environmental conditions.

Conclusions:

• AI-driven phenotyping and yield forecasting represent a paradigm shift in agricultural research.
• This integrated approach supports accurate and sustainable modern agriculture.
• The methodology enhances the ability to manage crops effectively in changing environments.