The detection of water stress in vineyards plays an integral role in the sustainability of high-quality grapes and prevention of devastating crop losses. Recent drought conditions in the Western Cape, seemingly also attributed to climate change, have added further urgency to this requirement. To model water stress in vineyards Stellenbosch University needed to sample the reflected spectral energy from the plant at a continuous range of spectral bands.
Images of water-stressed and non-stressed Shiraz vines were captured using the Simera Hyperspectral Imager. The imager captured 340 spectral bands across the visible and near-infrared (VNIR) range of 450-1000nm – wavebands with a bandwidth ranging from 0,9nm to 5nm. This data was combined with plant moisture measurement and machine-learning techniques to model and classify vineyard stress.
The combination of Simera’s Hyperspectral Imager and machine-learning techniques demonstrated the feasibility of creating a semi-automated framework for vineyard water stress modeling with test accuracies of above 80%. Refer to http://www.mdpi.com/2072-4292/10/2/202/htm for more information.