Machine Learning-Enabled Intelligent Technologies Across the Grape and Wine Value Chain: From Vineyard Sensing to...

Area of Science:

• Agricultural Science
• Computer Science
• Food Science

Background:

• Traditional grape and wine production relies on experience, but faces challenges in consistency and efficiency.
• The grape and wine industry is increasingly adopting data-driven approaches for improved decision-making.

Purpose of the Study:

• To synthesize machine learning (ML)-enabled technologies across the entire grape and wine value chain.
• To propose an integrated analytical framework for ML applications in viticulture and winemaking.
• To critically examine scientific challenges and emerging directions in ML for the wine industry.

Main Methods:

• Review of ML applications in vineyard sensing, precision viticulture, fermentation monitoring, and winemaking optimization.
• Analysis of multimodal data integration, model interpretability, and cross-domain generalization.
• Examination of sensing technologies (spectroscopy, chromatography-mass spectrometry, imaging, electronic sensing) and data preprocessing strategies.

Main Results:

• An end-to-end pipeline framework linking data acquisition to intelligent decision-making is proposed.
• Key challenges include multimodal data integration, model interpretability, generalization, decision coordination, and industrial deployment.
• ML offers significant potential for optimizing practices from vineyard management to quality evaluation.

Conclusions:

• Despite progress, challenges like data scarcity, heterogeneity, and implementation costs persist.
• Emerging ML approaches like knowledge-guided learning and causal inference are expected to improve robustness.
• This review provides a roadmap for advancing intelligent and sustainable practices in the grape and wine industry.