A spatio-temporal methodology for greenhouse microclimatic mapping.

Area of Science:

• Agricultural Engineering
• Environmental Science
• Data Science

Background:

• Greenhouse microclimates are often non-homogeneous, yet traditional models lack spatial resolution.
• Farmers require localized climate data for targeted interventions but face high sensor costs.

Purpose of the Study:

• To develop a data-driven methodology for creating virtual 2D greenhouse microclimate maps.
• To enable precise climate monitoring and control with minimal sensor investment.

Main Methods:

• Utilizing temporary sensors at key locations for data collection.
• Developing individual linear regression models for each sensor point.
• Employing spatial interpolation to generate a comprehensive 2D map.

Main Results:

• Achieved an adjusted R-squared value of 98% in predicting climate variables across different greenhouse locations.
• Demonstrated successful application in real-world greenhouse environments, monitoring temperature and other variables.

Conclusions:

• The proposed soft-sensing methodology offers an accurate and affordable solution for detailed greenhouse climate mapping.
• This approach empowers farmers with localized environmental data for optimized crop management and resource allocation.