80. How to best compare remote sensing data versus proximal sensing data?
Résumé
This work presents the application of a novel geostatistical method which, by identifying a unique optimal interpolation grid size, makes it possible to compare two spatial data sources with different measurement properties. The method relies on two steps. First it changes the spatial resolution of the data by identifying a common optimised interpolation grid that allows, for each data, a strong reduction in the impact of the stochastic variance while preserving the spatial variance. Second, it compares the data with linear regression modelling for gain and offset differences in the data. This approach was applied in the context of fertilisation recommendations in viticulture, making it possible to integrate the advantages of airborne NDVI measurements (larger scale, lower cost, etc.) into an advisory process developed for vehicle-mounted (proximal,) NDVI measurements.