Functional unfold principal component analysis for automatic plant-based stress detection in grapevine.

Area of Science:

• Agricultural Science
• Plant Physiology
• Data Mining

Background:

• Grapevine water status significantly impacts grape and wine quality, making drought stress detection vital.
• Stem diameter variations can indicate drought stress but are influenced by microclimatic changes.
• Advanced data analysis is needed to differentiate drought effects from microclimate variations.

Purpose of the Study:

• To evaluate the effectiveness of Unfold Principal Component Analysis (UPCA) and Functional Unfold Principal Component Analysis (FUPCA) for detecting drought stress in grapevines.
• To compare the performance of FUPCA against UPCA in identifying drought-induced deviations in stem diameter variations.

Main Methods:

• Applied UPCA, an established data mining technique, to analyze grapevine stem diameter variations.
• Utilized FUPCA, a novel method combining functional data analysis with UPCA, to process multivariate time series data.
• FUPCA involves approximating time series with basis functions and applying UPCA to the resulting coefficients.

Main Results:

• Both UPCA and FUPCA successfully detected drought stress in grapevines days before the appearance of visible symptoms.
• FUPCA demonstrated greater robustness and statistical simplicity compared to UPCA.
• FUPCA exhibited the capability to handle datasets with missing data points, a limitation of UPCA.

Conclusions:

• FUPCA is a superior tool for detecting drought stress in grapevines, offering improved accuracy and data handling capabilities.
• The findings suggest FUPCA can aid viticulturists in proactive drought management, preserving grape and wine quality.
• This study highlights the potential of advanced functional data analysis techniques in precision agriculture.