ISHS Acta Horticulturae 1390: XI International Symposium on Grapevine Physiology and Biotechnology
Relationship between vegetation indices calculated from Sentinel-2 imagery and vineyard ground-truth data |
| Authors: | S. Vélez, J.A. Rubio, E. Barajas |
| Keywords: | vegetation indices, vegetative and generative parameters, NDVI, remote sensing, Sentinel |
| DOI: | 10.17660/ActaHortic.2024.1390.37 |
| Abstract:
Remote sensing is a tool that has been increasingly used to obtain important information about crops. This information is captured by sensors mounted on several platforms, such as ground vehicles, airborne or even satellites. Each platform has its advantages and disadvantages, but satellites such as Sentinel-2 can be especially useful. In addition, the use of satellite images has proven to be useful in viticulture. A typical way to use this information is calculating vegetation indices (VIs), algebraic combinations designed to highlight the contrast of specific properties or plant characteristics. One of the most used indices is NDVI (normalized difference vegetation index), typically computed in veraison. However, other indices may be used in vineyards with good results, such as SAVI (soil adjusted vegetation index), CTVI (corrected transformed vegetation index) or NDRE (normalised difference red edge index). This paper aims to analyse the relationship between the most commonly used vegetation indices, derived from Sentinel-2 images, and the ground-truth data measured in the vineyard, to assess the relevancy of each index and the best time of year to calculate them. The results show CTVI and EVI (enhanced vegetation index) are more correlated with vegetative parameters and mcari (modified chlorophyll absorption ratio index) with generative parameters. Moreover, veraison is an appropriate phenological stage to compute them, but other growing stages also show to be relevant with better correlation. | |
Download Adobe Acrobat Reader (free software to read PDF files) | |
