Coupling machine vision and crop models for closed-loop plant production in advanced life support systems

Area of Science:

• Controlled environment agriculture
• Plant science
• Biotechnology

Background:

• Advanced life support systems require efficient food production methods.
• Integrating real-time monitoring with predictive models is crucial for optimizing plant growth in controlled environments.
• Nondestructive sensing offers a non-invasive approach to gather plant growth data.

Purpose of the Study:

• To present a conceptual framework for coupling nondestructive sensing technologies with crop growth models.
• To enable closed-loop plant production systems for applications such as NASA's advanced life support.
• To utilize plant growth data for system diagnostics and model-based feedback.

Main Methods:

• Developed a framework linking nondestructive observations to crop model predictions.
• Utilized canopy height and top projected canopy area (TPCA) measured by machine vision.
• Employed the CROPGRO crop growth model for simulations.
• Validated model predictions against hydroponic soybean data under varied temperature regimes.

Main Results:

• Soybean top projected canopy area (TPCA) and canopy height were simulated using the CROPGRO model.
• Simulations were compared with experimental data from hydroponic soybean grown at two different temperature conditions (23/19°C and 26/22°C).
• Canopy height and TPCA demonstrated potential as useful variables for closed-loop systems.

Conclusions:

• Nondestructive sensing coupled with crop models provides valuable feedback for plant production systems.
• Top projected canopy area (TPCA) and canopy height are identified as key variables for early-stage closed-loop plant production.
• This approach is particularly relevant for controlled environment agriculture and space exploration.