Investigating a Selection of Methods for the Prediction of Total Soluble Solids Among Wine Grape Quality
Aikaterini Kasimati1, Borja Espejo-Garcia1, Eleanna Vali1
1Natural Resources Management and Agricultural Engineering, Agricultural University of Athens, Athens, Greece.
Frontiers in Plant Science
|June 28, 2021
Summary
Predicting wine grape quality using machine learning and NDVI data offers a faster, non-destructive alternative to lab analysis. Unmanned aerial vehicle (UAV) and proximal sensors show promise for accurate quality prediction.
Area of Science:
- Agricultural Science
- Remote Sensing
- Machine Learning
Background:
- Traditional wine grape quality assessment relies on time-consuming and costly laboratory analysis of samples.
- Precision viticulture (PV) utilizes non-destructive sensing methods, including proximal and remote sensing, for vineyard management.
- Normalized Difference Vegetation Index (NDVI) is a key vegetation index used to assess plant health and characteristics.
Purpose of the Study:
- To investigate an alternative, non-destructive approach for predicting wine grape quality characteristics.
- To combine machine learning techniques with NDVI data from various sensing platforms.
- To evaluate the performance of proximal and remote sensing data at different grapevine growth stages.
Main Methods:
- Collected high-resolution multispectral data from vehicle-mounted sensors, unmanned aerial vehicles (UAVs), and Sentinel-2 (S2) imagery.
- Pre-processed data including quality assessment, interpolation, and normalization.
- Applied statistical analysis (Pearson's correlation) and machine learning regression models (OLS, Theil-Sen, Huber, Decision Trees) to predict grape quality.
Main Results:
- Proximal sensors were more effective for early-season quality prediction, while remote sensors performed better in later stages.
- Strongest correlations between NDVI and sugar content were observed using UAV and Spectrosense+GPS (SS) data during the 'Berries pea-sized' and 'Veraison' stages.
- The best regression models achieved a maximum coefficient of determination (R²) of 0.61.
Conclusions:
- Combining machine learning and NDVI data provides a viable non-destructive method for predicting wine grape quality.
- UAV and SS data show high accuracy in predicting sugar content, particularly in mid-to-late season growth stages.
- This approach can significantly improve the efficiency and reduce the cost of wine grape quality assessment.
Related Concept Videos
02:35Key Elements for Plant Nutrition
22.9K
Like all living organisms, plants require organic and inorganic nutrients to survive, reproduce, grow and maintain homeostasis. To identify nutrients that are essential for plant functioning, researchers have leveraged a technique called hydroponics. In hydroponic culture systems, plants are grown—without soil—in water-based solutions containing nutrients. At least 17 nutrients have been identified as essential elements required by plants. Plants acquire these elements from the...
22.9K
01:19Precipitation Titration: Endpoint Detection Methods
2.4K
In argentometric precipitation titrations, endpoints can be detected visually by the Mohr, Volhard, and Fajans methods. In the Mohr method, adding a soluble chromate indicator gives an initial yellow color to the analyte solution. As the titrant is added, the first excess of silver ions forms a red silver chromate precipitate, marking the endpoint. The solution pH should be maintained at about 8 by adding solid CaCO3.
In the Volhard method, a standard excess of AgNO3 is first added to the...
In the Volhard method, a standard excess of AgNO3 is first added to the...
2.4K