Fungal pathogens constitute the greatest economical concern to corn farmers; they result in yield losses, grain quality reduction and production of mycotoxins. Improvement of detection methods are imperative. This work aimed to examine corn fungal pathogens with HSI.
Isolates of Fusarium spp. and Stenocarpella spp., were plated on growth media in glass Petri dishes in triplicate, and incubated at 25°C for 9 days. Images were acquired with a SisuChema short-wave infrared pushbroom imaging system in the spectral range 920 – 2514 nm. Principal component analysis (PCA), with various pre-processing methods, and multivariate curve resolution (MCR) were used to explore the data.
PCA with or without pre-processing, revealed chemical differences within and between fungal isolates. Differences were amplified with time. Examination of the mean spectra and PC loadings after spectral pre-treatment indicated variation primarily around bands associated with water/moisture (1450 & 1930 nm), protein (2180 & 2242 nm) and carbohydrates/starch (1090, 1360 & 2100 nm). This is expected since fungi are mainly comprised of these constituents and as the mycelium grows and ages, there is a change in carbohydrate (content or structure), moisture and protein. This was apparent in higher order components (PCs 4-6) and appeared as textured information.
MCR revealed similar results, however the concentration maps were clearer than PCA score images. In addition, these maps were textured illustrating the physical changes of the mycelium with time. These were due to the growing hyphae and possible spore formation. In addition, it is likely that these concentration maps indicated presence of mycotoxins.