Detection of significant wavelengths for identifying and classifying Fusarium oxysporum during the incubation period and water stress in Solanum lycopersicum plants using reflectance spectroscopy

Abstract

Spectroscopy has become one of the most used non-invasive methods to detect plant diseases before symptoms are visible. In this study it was possible to characterize the spectral variation in leaves of Solanum lycopersicum L. infected with Fusarium oxysporum during the incubation period. It was also possible to identify the relevant specific wavelengths in the range of 380–1000 nm that can be used as spectral signatures for the detection and discrimination of vascular wilt in S. lycopersicum. It was observed that inoculated tomato plants increased their reflectance in the visible range (Vis) and decreased slowly in the near infrared range (NIR) measured during incubation, showing marked differences with plants subjected to water stress in the Vis/NIR. Additionally, three ranges were found in the spectrum related to infection by F. oxysporum (510–520 nm, 650–670 nm, 700–750 nm). Linear discriminant models on spectral reflectance data were able to differentiate between tomato varieties inoculated with F. oxysporum from healthy ones with accuracies higher than 70% 9 days after inoculation. The results showed the potential of reflectance spectroscopy to discriminate plants inoculated with F. oxysporum from healthy ones as well as those subjected to water stress in the incubation period of the disease.

Keywords

• band selection
• hyperspectral reflectance
• plant diseases
• relevant specific wavelength
• vascular wilt