Phytopathologia Mediterranea (Dec 2018)

Detection of grapevine leaf stripe disease symptoms by hyperspectral sensor

  • Amanda JUNGES,
  • Jorge DUCATI,
  • Cristian SCALVI LAMPUGNANI,
  • Marcus André ALMANÇA

DOI
https://doi.org/10.14601/Phytopathol_Mediterr-22862
Journal volume & issue
Vol. 57, no. 3

Abstract

Read online

Hyperspectral sensors can measure reflectance in a wide range of the electromagnetic spectrum. These can be used as an indirect method for detecting plant disease, by comparing the specific spectral signatures between symptomatic and asymptomatic vegetation. Grapevine Leaf Stripe Disease (GLSD), including the Esca complex, is a very important Grapevine Trunk Disease (GTD) worldwide. With the objective of developing an innovative method for quantitative and qualitative analyses of symptomatic plants using remote sensing, this study measured and characterized the spectral behaviour of GLSD asymptomatic and symptomatic grapevine leaves using a hyperspectral sensor. Asymptomatic, initial and final GLSD symptomatic leaves were collected in two stages of the phenological cycle (before and after harvest) from a ‘Merlot’ vineyard in Veranópolis, Rio Grande do Sul, Brazil. Reflectance measurements (350 to 2,500 nm) were performed using a spectroradiometer. The spectral behaviour of vine leaves with GLSD symptoms changed especially in the visible light range; reflectance increased in the green edge (520-550 nm) and red edge (700 nm) associated with reduced photosynthetic pigments (especially chlorophyll b). At near-infrared, reflectance decreased, especially in leaves with advanced GLSD, due to cell structure loss and toxin accumulation induced by pathogens. Even at different intensities, leaf reflectance changed in initial and final GLSD symptoms and at different stages of the cycle. These results showed that proximal, non-destructive sensing techniques may be useful tools for detecting the changing spectral behaviour of grapevine leaves with GLSD, which could be used for disease identification and detection.

Keywords