Advanced phenotyping platform for rice breeding under water-limited conditions

Overview

This solution offers a sensor-based phenotyping platform designed to accelerate rice breeding for water-limited, non-flooded cultivation systems. As global water scarcity intensifies, productive rice varieties that thrive under irrigated (non-flooded) conditions are increasingly needed. The platform integrates proximal and remote sensing techniques—including RGB, multispectral, and thermal imagery, together with hyperspectral reflectance of plants and grains—to improve genotype selection in rice breeding programs. By combining field and aerial data acquisition with predictive modeling, the approach streamlines the evaluation of large genotype panels for both agronomic performance and grain quality.

Technical specifications

Imaging and sensing modalities:

• Aerial and field RGB, multispectral, and thermal imagery for canopy-level assessment
• Hyperspectral reflectance measurements on plants and grains
• Canopy temperature used to calculate stress degree day (SDD) indices

Validated predictive models developed by the research group:

• Spectral reflectance plus support vector regression models for wheat predicting grain yield (90% variance explained), carbon isotopic discrimination (92%), and spikes per m² (73%)
• Significant correlation between SDD and grain yield (r² = 0.66)
• Random forest regression models on rice grain hyperspectral data predicting chalkiness (R² = 0.98), translucency (R² = 0.84), and milling degree (R² = 0.88)

Application areas:

• High-throughput genotype selection in rice breeding programs targeting water-limited environments
• Evaluation of agronomic, physiological, and grain quality traits
• Scalable phenotyping across large genotype panels (e.g., 384 wheat genotypes and 310 rice grain samples validated to date)

Technology readiness level

The underlying phenotyping methods and predictive models have been validated in wheat, triticale, alfalfa, and fruit crops, with recent extension to rice grain quality traits. A one-year validation study is planned in which 40 advanced rice lines will be grown under irrigation and evaluated using aerial and field imagery, agronomic trait monitoring, and grain quality analysis. The team is actively seeking funding for equipment acquisition and key personnel support to advance this next validation phase. The platform is positioned for collaborative development with breeding programs and industry partners seeking to deploy sensor-based selection tools for rice under water-scarce conditions.