Precision Agriculture Using Soil Sensor Driven Machine Learning for Smart Strawberry Production

Area of Science:

• Agricultural Engineering
• Precision Agriculture
• Sensor Networks

Background:

• Industrial settings increasingly use ubiquitous sensor networks for real-time data collection.
• Precise monitoring and control of soil conditions are crucial for optimizing crop production.

Purpose of the Study:

• To describe an end-to-end system for precise monitoring and control of soil conditions in strawberry farming.
• To infer physicochemical characteristics of strawberries at harvest using sensor data.
• To develop predictive models for key strawberry quality attributes.

Main Methods:

• Utilized a sensor network distributed in the soil of a commercial strawberry farm.
• Employed empirical and statistical models, including neural networks and Gaussian process regression.
• Investigated the prediction of physicochemical qualities such as color and soluble solids content.

Main Results:

• Accurate prediction of strawberry color within 9% of expected values.
• Prediction of combined color and soluble solids content (sweetness) within 14% of expected values.
• Demonstrated the potential for data-driven control of soil conditions.

Conclusions:

• The developed system enables precise monitoring and prediction of strawberry quality.
• High prediction accuracy facilitates data-driven decisions for sustainable and high-quality strawberry production.
• The framework supports resource-use optimization and quality-resource trade-offs in agriculture.