Automatic detection of powdery mildew in grapevine: imaging approachs for accurate sensing in field conditions

Powdery mildew is a major fungal disease for grapevine, and for other specialty crops as well as, causing severe damagesto yield and quality of the produce. This disease is thoroughly controlled by uniform fungicides spraying to vineyards,nevertheless its primary infections emerge from uneven discrete foci. There is then an evident potential of benefitsassociated to the development technologies for high-precision crop protection, i.e. systems able to detect initial infection foci and to operate targeted treatments on them, instead of applying homogenous and unselective sprayings as currently done. Proximal optical sensing from tractor, or other field platform,is a major candidate technique to early detect infection foci in grapevine and other specialty crops with vertical canopy structure.In the case of powdery mildew infection, anyway, the sensitivity in detecting early symptoms can be largely limited by the combination of small dimensions, low density, and spatial arrangement of thin fungal structures. This paper illustrates some of the results obtained by theauthors in different experiments conducted on grapevine‘s powdery mildew automatic detection by means of multispectral and hyperspectral imaging, and it discusses the measurements approaches aimed to improve the accuracy of the detection in field conditions and the data analysis algorithms.