A FFT-BASED APPROACH TO EXPLORE PERIODICITY OF VINES/SOIL PROPERTIES IN VINEYARD FROM TIME SERIES OF SATELLITE-DERIVED SPECTRAL INDICES

In the Precision Viticulture context, NDVI already proved to be correlated to vigour and midday stem water potential of vines; NDWI to soil volume water content. It is thus expected that significant periodicities can be found looking at spectral indices time series. This can be useful to better interpret vines behaviour and, possibly, relate intra-vineyard macroscopic differences to this type of information. To preliminarily test this hypothesis 25 Landsat 8 OLI images, Level-2 Data Products, were processed to compute NDVI and NDWI time series. An interpolation step aimed at generating daily estimates of indices in the explored period (May 2013-February 2016) was performed. Successively vegetation effects were removed from NDWI to emphasize soil signal. Finally a FFT analysis of both NDVI and NDWI interpolated time series was achieved to explore their periodicity, giving a first agronomic interpretation of results.