Hyperspectral imaging platform for early detection of crown rot disease in wheat

Overview

Crown rot is a destructive fungal disease of wheat that damages the vasculature of individual tillers, restricting water and nutrient uptake and causing premature plant death. Traditional screening relies on visual human scoring, which is time-consuming, subjective, and difficult to scale across breeding programmes and field trials. This solution proposes a hyperspectral imaging-based method that detects the biochemical and physiological changes in wheat shoots caused by crown rot, enabling rapid, quantitative, and cost-effective disease assessment directly in field conditions.

The approach addresses a critical bottleneck for cereal breeders, agronomists, and crop protection companies seeking to accelerate the identification of resistant cultivars and improve disease management decisions.

Technical specifications

• Spectral range: Hyperspectral imaging across visible and infrared bands from 400 to 2,500 nm, capturing subtle biochemical and physiological signatures in wheat shoots.
• Sensor platform: A ground-based FieldExplorer phenotyping platform designed to collect high-resolution hyperspectral imagery under field conditions.
• Detection targets: Altered nitrogen content, shoot water content, and biochemical defence pathway responses triggered by crown rot infection and the resulting vasculature restriction.
• Analytical approach: Classification and regression models trained to quantify disease severity from hyperspectral data, building on validated methods that estimate nitrogen content with a regression accuracy of R² = 0.88.
• Experimental design: Manipulation of Genotype × Environment × Management (G×E×M) interactions across multiple cultivars and treatments to develop a robust, broadly applicable prediction model.
• Validation: Model outputs are validated by plant pathology experts to ensure reliability for breeding and crop management applications.

Technology readiness level

The research builds on demonstrated laboratory capability, including accurate estimation of nitrogen content in wheat shoots from hyperspectral data (R² = 0.88) and prior evidence that shoot water content changes can be detected spectrally. A funded greenhouse-based pilot experiment is already underway. The next stage is field validation using the FieldExplorer platform across diverse cultivars and management treatments to refine the prediction model. The technology is currently at an early-to-mid development stage, progressing from controlled-environment validation toward field-ready deployment for breeding and agronomic screening.