Short-wave infrared (SWIR) spectral analysis of healthy and Erwinia amylovora inoculated apple trees, the base of a new machine vision system for precision farming

Fire blight, caused by the bacterium Erwinia amylovora , poses a significant economic threat to global apple production due to its rapid spread and destructive nature. This study evaluates the efficacy of Short-Wave Infrared (SWIR) sensing, utilizing both point-based Micro-Electro-Mechanical Systems (MEMS) spectroscopy and hyperspectral imaging (HSI), for the early detection of infections in ‘Najdared’ apple trees. Our results demonstrate that Partial Least Squares (PLS) regression of the point-based SWIR spectra captured spectral changes associated with disease progression, achieving a coefficient of determination ( R 2 ) of up to 0.99 in internal cross-validation. Most notably, the HSI-based classification provided a five-to-six-day lead time in symptom detection compared to traditional visual inspection. These findings support the feasibility of SWIR-based machine vision for proactive orchard management and the development of automated, cost-effective monitoring systems.