Comparison Between Different Spatial Resolution Satellites Images For Viticulture

Knowledge of spatial variation of vegetative vigouris essential for site-specific management and variable rate application. Specifically, precision viticulture requires continuous monitoring of the vigour due to the mutable condition during the year and to frequent treatments (10 to 15) during the growingseason. Over the last few years, several solutions for vegetation variability surveying have been proposed, mainly based on optical remote and proximal sensing. In precision viticulture, images fromremote sensingare influencedby inter-row spacing, which can somehow mask or distort field variability. Analysis of high-resolution satellites or UAV imagery are not trivial, due to the computing power needs, howeverinter-row spacing can be in these cases recognized, segmented and eventually removed. On the other hand, medium resolution satellites imagery does not allow identification of single rows, while they can easilyprovide general averaged information. The spatial resolution of the images plays a preeminent role in vineyard variability sensing and understanding. In the present work, vineyard variability is evaluatedthrough vegetation indices, which were estimated taking advantage of multispectral images at different resolutions, from Planet Scope, RapidEye and Sentinel 2 satellites respectively providing 3, 5 and 10 meters pixel sizes. The experimentation took place in a 5hectares vineyard located in Northern part of Italy. Data were comparedwith UAV images, opportunely re-sampled in order tofit the spatial resolution of different considered satellites. The comparison wascarried out considering referencesstatistical parameters calculated on the basis ofspectral images. Fast Fouriertransform,and root mean squarewasbeen considered to exploit differences and relation between sources images.