Traits Expansion and Storage of Soybean Phenotypic Data in Computer Vision-Based Test

Area of Science:

• Agricultural Science
• Plant Breeding
• Computer Vision

Background:

• Crop phenotypic traits are crucial for developing new varieties.
• Current trait evaluation methods are labor-intensive and may not suit machine learning.
• Artificial intelligence (AI) and computer vision (CV) offer potential for efficient, accurate phenotyping.

Purpose of the Study:

• To expand the scope of phenotypic traits for soybean variety testing using CV.
• To demonstrate the necessity and benefits of incorporating new traits for AI-driven breeding.
• To enhance data collection for AI Soybean Breeding Platforms.

Main Methods:

• Utilized computer vision (CV) technology for phenotypic data collection.
• Expanded trait categories beyond existing manual examination criteria.
• Focused on four key soybean trait types: field agronomic, indoor agronomic, resistance, and seed traits.

Main Results:

• Successfully identified and incorporated new phenotypic traits suitable for CV analysis.
• Demonstrated the inadequacy of current criteria for machine learning applications.
• Highlighted the value of expanded traits for AI-based breeding programs.

Conclusions:

• The expanded phenotypic traits are essential improvements to existing soybean testing guidelines.
• CV-based trait data collection supports the development of AI Soybean Breeding Platforms.
• These advancements facilitate accurate breeding program design and accelerate crop improvement.