High-Throughput Imaging System for Early Detection of Root Rot Disease in Field Peas (Pisum sativum L.)

Aphanomyces root rot (ARR) and Fusarium root rot (FRR), caused by Aphanomyces euteiches and Fusarium avenaceum, are major soil-borne diseases limiting field pea productivity and breeding progress. This study presents a high-throughput hyperspectral imaging (HSI) system combined with machine learning (ML) to enable early, non-destructive detection of root rot in controlled greenhouse trials. Sixteen pea genotypes were evaluated across three experiments in 2024, with shoot and root reflectance data acquired using a HSI sensor. Visual disease ratings were categorized into three severity levels of low, moderate, and high for model development. ML algorithms, including Partial Least Squares Regression (PLSR), Random Forest (RF), and Artificial Neural Network (ANN), were trained to predict disease severity. PLSR achieved the highest predictive accuracy, especially when using root spectra, which showed best performance in detecting high disease severity. Key wavelengths in the red-edge (680–750 nm) and near-infrared (750–1000 nm) regions were identified as critical for early disease inference. Preliminary relationships linking shoot and root spectra were also explored for field applicability. This integrated HSI-ML approach offers a scalable, rapid phenotyping tool for precision breeding, supporting the development of disease-resistant pea cultivars and advancing sustainable crop protection strategies.