An UAV-based remote sensing approach for the detection of spatial distribution and development of a grapevine trunk disease

Abstract

The availability of new tools to detect and monitor vines health status is under evaluation in a vineyard heavily affected by grapevine leaf stripe disease (GLSD). The vineyard, located in Tuscany, has been surveyed for foliar symptoms each year since 2003, and each vine was mapped and classified for foliar symptoms appearance and diseases severity. At the beginning of September 2011 an UAV (Unmanned Aerial Vehicle) system was used to investigate the correlation between Normalised Differential Vegetation Index (NDVI) and GLSD foliar symptoms. The acquired high resolution multi-spectral images (0.056 m/pixel) were also analyzed for the detection of vines that had shown symptoms in the previous years, but were asymptomatic in 2011. All sample vines locations were accurately measured with a differential GPS to extract single plant reflectance from the georeferenced multispectral images. Moreover during the following winter the vigor of the surveyed vines was measured by cane lengths, number of canes produced and total dry mass weight. The preliminary results obtained in the pilot experiment, showed a high level of correlation between NDVI and GLSD symptoms. The method has demonstrated both the ability to discriminate symptomatic from asymptomatic plants, but also to detect within the asymptomatic plants sampled, those which were asymptomatic but had shown GLSD symptoms in the previous years, suggesting the possible developments as an innovative early detection tool for GLSD. The work offers a significant potential both in terms of research perspectives and practical applications to improve the mapping of spatial distribution and development of GLSD.