Application of Parallel Factor Analysis (PARAFAC) to the Regional Classification of Vineyard Blocks Using Remote Sensing
Eva Lopez-Fornieles,* Guilhem Brunel, Nicolas
Devaux, Jean-Michel Roger, James Taylor, and Bruno Tisseyre
*ITAP, University of Montpellier, Institut Agro, INRAE,
Montpellier, 34000, France (eva. fornieles-lopez@supagro.fr)
Monitoring winegrowing regions and maximizing the value of production based on regional and local specificities requires accurate spatial and temporal monitoring. The increasing amount and variability of information from remote sensing data is a potential tool to assess this challenge in the grape and wine industry. We examined the capacity of a multiway analysis method applied to the Sentinel-2 time-series to assess the value of simultaneously considering spectral and temporal information to differentiate production differences at a regional scale. PARAllel FACtor analysis (PARAFAC) was used as an unsupervised technique to recover pure spectra from multi-way spectral imagery of vineyards in the Languedoc-Roussillon region in the south of France. The model was developed using a time series of Sentinel-2 satellite imagery collected on 4978 vineyard blocks between May 2020 and June 2021. The Sentinel-2 signal was resolved into spectral and temporal profiles in the form of pure compounds, which are potentially specific to both vegetation and soil. The PARAFAC analysis indentified two pure compounds strongly related to characteristics and dynamics of vineyard cultivation on a regional scale. A conceptual framework was proposed using the simultaneous integration of spectral and temporal attributes from Sentinel-2 time-series at regional scale with the PARAFAC methodology and was validated with a practical framework of expert winegrowers’ opinions from the Languedoc Roussillon region. The introduction of PARAFAC insights into the analysis approach, from a conventional grapegrower’s perspective, opens the possibility to identify spectro-temporal profiles of vineyard blocks relevant for understanding and characterizing them on a regional scale.
Funding Support: This work was supported by the French National Research Agency under the Investments for the Future Program, referred as ANR-16-CONV-0004.
