Uas-based disease diagnostics and yield prediction for cotton and other crops

Overview

This solution leverages unmanned aerial system (UAS) imaging combined with advanced data analytics to detect crop diseases, assess their severity, and predict yield losses before harvest. Originally validated in cotton fields affected by Fusarium wilt race 4, the approach correlates canopy-level image data with actual yield outcomes, achieving a 0.96 correlation in preliminary studies. The technology offers growers, breeders, and crop protection companies a scalable, non-destructive tool for early disease detection, variety evaluation, and precision management decisions.

Beyond disease diagnostics, the platform can be applied to screen germplasm for environmental tolerances such as drought, making it valuable for plant breeding programs and agronomic research. By mapping affected areas within a field, the technology supports site-specific interventions, including targeted chemical applications and deployment of resistant genotypes, reducing input costs and environmental impact.

Technical specifications

Imaging and data collection:

• UAS platforms capture high-resolution aerial imagery over crop fields
• Phenotypic data collected includes canopy cover, canopy height, and canopy length
• Disease response and environmental data are recorded alongside imaging data
• Validated in cotton variety trials under Fusarium wilt race 4 infestation pressure

Analytical methods:

• Structure from Motion (SfM) techniques generate three-dimensional crop models
• Digital Surface Models (DSM) quantify canopy structure and biomass indicators
• Binary algorithms analyze image data to detect disease-affected areas and predict yield potential
• Correlation analysis links canopy metrics to harvested yield with demonstrated accuracy of 0.96

Key features:

• Early disease detection before visible symptoms appear at ground level
• Severity assessment based on canopy impact and crop establishment
• Mapping of affected zones for precision management
• Germplasm screening capability for traits such as drought tolerance
• Scalable across field sizes and adaptable to additional crops and pathogens

Technology readiness level

The technology has been validated through field trials in cotton at El Paso, Texas, where UAS-collected canopy cover data correlated with yield at 0.96. Current efforts focus on expanding validation in additional fields infested with Fusarium wilt, incorporating variety trials and broader phenotypic and environmental datasets. Future work aims to refine disease severity detection algorithms and demonstrate site-specific management applications. The platform is positioned for collaborative refinement and pilot-scale deployment with industry partners in agriculture, crop protection, and plant breeding.