Abstract

Pitch canker, caused by Fusarium circinatum, poses a major threat to Pinus radiata plantations, resulting in substantial economic and ecological losses. Early detection of this pathogen is crucial, as conventional methods rely on late-stage visual symptoms. This study explores the potential of visible-near-infrared hyperspectral imaging (VIS-NIR HSI) combined with multivariate techniques for the early detection of F. circinatum infection in P. radiata cuttings before symptom onset. The infection process was monitored over 57 days in two P. radiata genotypes through hyperspectral image acquisition in the 400-1000 nm range. Fast Principal Component Analysis (Fast-PCA) and Partial Least Squares Discriminant Analysis (PLS-DA) were applied to identify key spectral variations and classify samples as infected or healthy, respectively. The results demonstrate that early differentiation between infected and healthy cuttings is possible, achieving high classification accuracy at 27 days post-inoculation (dpi) in predictive model validation. Additionally, phenotypic differences between genotypes were observed, with genotype A exhibiting earlier and more pronounced spectral changes between infected and control samples than genotype B, suggesting varying resistance levels of genotypes. These findings underscore the potential of VIS-NIR HSI for both early disease detection and the assessment of genetic susceptibility, providing valuable insights for breeding programs aimed at enhancing P. radiata resistance while establishing HSI as a powerful, non-invasive, and high-throughput phenotyping tool with applications in precision forestry and large-scale disease monitoring.