terclim by ICS banner
IVES 9 IVES Conference Series 9 GiESCO 9 Deep learning based models for grapevine phenology

Deep learning based models for grapevine phenology

Abstract

Context and purpose of the study – the phenological evolution is a crucial aspect of grapevine growth and development. Accurate detection of phenological stages can improve vineyard management, leading to better crop yield and quality traits. However, traditional methods of phenological tracking such as on-site observations are time-consuming and labour-intensive. This work proposes a scalable data-driven method to automatically detect key phenological stages of grapevines using satellite data. Our approach applies to vast areas because it solely relies on open and satellite data having global coverage without requiring any in-field data from weather stations or other sensors making the approach extensible to other areas.

Material and methods we leveraged historical phenological observations and developed a supervised deep-learning model that uses the land surface temperature estimated by the Copernicus Sentinel-3 satellite to estimate the current phenological stage at the parcel level. We compared the performances of our model with traditional approach based on Growing Degrees Days (GDD).

Results – we train our algorithm on manually collected phenological observations of four winegrape cultivars in three Europeanvineyards (Italy, Spain, and Portugal) from 2017 to 2022. Preliminary results indicated that our deep learning phenology model outperforms the traditional methods based on GDD, decreasing the Mean Absolute Error from 33.8 to 7.8 days (-76.5%).

DOI:

Publication date: July 5, 2023

Issue: GiESCO 2023

Type: Poster

Authors

Federico OLDANI1*, Dario SALZA1, Giacomo BLANCO1, Claudio ROSSI1, Boris BASILE2*, Fabrizio CARTENI2, Núria PÉREZ-ZANÓN3, Antonio DENTE4, Fernando ALVES5, Joana VALENTE5, Montse TORRES6, Carlos EZQUERRA6, Rosa ARAUJO7

1LINKS Foundation, Turin, Italy
2Department of Agricultural Sciences, University of Naples Federico II, Portici (Napoli), Italy
3Barcelona Supercomputing Center, Barcelona, Spain
4Mastroberardino, Atripalda (Avellino), Italy
5Symington Family Estates, V. N. Gaia, Portugal
6Familia Torres, Vilafranca del Penedès, Spain
7Eurecat, Centre Tecnològic de Catalunya, Barcelona, Spain

Contact the author*

Keywords

satellite imagery, earth observation, machine learning, Sentinel-3, Copernicus, climate change

Tags

GiESCO | GIESCO 2023 | IVES Conference Series

Citation

Related articles…

Quantification of newly identified C8 aroma compounds in musts and wines as an analytical tool for the early detection of Fresh Mushroom Off-Flavor

The Fresh Mushroom Off-Flavor (FMOff) is a concerning undesirable aroma in wine specific of certain vintages, characterized by a typical button mushroom aroma. The appearance of this off-flavor is linked to the presence of certain fungus on the grape [1-3].

Comparison of the principal production methods for alcohol-free wine based on analytical parameters

Production, demand, and brand awareness of dealcoholized wine (<0.5% v/v) is steadily increasing worldwide. However, there have been few studies to date investigating and comparing the different physical processes for dealcoholizing wine.

Evolution of oak barrels C-glucosidic ellagitannins in model wine solution

Oak wood has a significant impact on the chemical composition of wine, leading to transformations that influence its organoleptic properties, such as its aroma, structure, astringency, bitterness and color. Among the main extractible non-volatile polyphenol compounds released from oak wood, the ellagitannins are found [1].

Single plant oenotyping: a novel approach to better understand the impact of drought on red wine quality in Vitis x Muscadinia genotypes

Adopting disease-tolerant varieties is an efficient solution to limit environmental impacts linked to pesticide use in viticulture. In most breeding programs, these varieties are selected depending on their abilities to tolerate diseases, but little is known about their behaviour in response to abiotic constraints.

Quality assessment of partially dealcoholized and dealcoholized red, rosé, and white wines: physicochemical, color, volatile, and sensory insights

The global non-alcoholic wine market is projected to grow from USD 2.7 billion in 2024 to USD 6.97 billion by 2034, driven by health awareness, lifestyle shifts, and religious factors [1-3]. Consequently, the removal of alcohol can significantly alter the key quality parameters of wine.