Fusion of IR and Multispectral Images in the Visible Range for Empirical and Model Based Mapping of Crop Water Status

Abstract

Achieving high quality wine grapes depends on the ability to maintain mild stress in the crop during the growing season. Natural geographical variability can lead to variable stress levels in different parts of the vineyard resulting in variability in grape quality at harvest. This can be prevented by precision irrigation at the vine level. Therefore, the ability to monitor stress levels at the plant and plot scale can increase orchard profitability by facilitating the production of a uniform crop of high quality grapes. The current research is part of an ongoing investigation of the use of thermal imaging for monitoring crop water stress. Three irrigation treatments were applied in a wine-grape (vitis vinifera cv. Merlot) vineyard in northern Israel. Thermal and visible (RGB) images of the crop were and crop parameters, as well as climate parameters were monitored. Crop water stress index (CWSI) was calculated using a variety of leaf, canopy, and reference temperatures determined from the thermal images and climate data. The overlay of RGB and thermal images allowed selection of sunlit leaf temperatures. CWSI was highly correlated with both stomatal conductance and stem water potential. The relationship to stomatal conductance did not change significantly during the season, but for that to stem water potential both intercept and slope varied. The highest coefficient of correlation was obtained when CWSI was computed using sunlit leaf temperatures from the center of the row, reference wet temperature from the artificial wet surface, and dry leaf temperature reference from air temperature plus 5°C.