Non-Destructive Monitoring of Crop Fresh Weight and Leaf Area with a Simple Formula and a Convolutional Neural...

Area of Science:

• Agricultural Science
• Horticulture
• Plant Physiology

Background:

• Non-destructive growth factor measurement is crucial for crop monitoring.
• Existing methods for measuring crop fresh weight and leaf area can be complex and limited.
• A versatile, non-destructive method is needed for efficient crop assessment.

Purpose of the Study:

• To establish a non-destructive monitoring system for estimating fresh weight and leaf area in trellised crops.
• To develop accurate and versatile methods for real-time crop growth assessment.
• To provide a practical solution for data collection in agricultural research.

Main Methods:

• Fresh weight estimation using total system weight and volumetric water content via a simple formula.
• Leaf area estimation employing top-view images and a convolutional neural network (ConvNet).
• Data collection from sweet pepper (Capsicum annuum) crops in a greenhouse environment.

Main Results:

• The monitoring system achieved an average R² of 0.70 for fresh weight estimation.
• The system demonstrated a high average R² of 0.95 for leaf area estimation using ConvNet.
• The simple calculation method avoided overfitting and showed fewer limitations than previous approaches.

Conclusions:

• The developed monitoring system effectively estimates crop fresh weight and leaf area using non-destructive methods.
• The combination of simple calculations and ConvNet offers a versatile and accurate approach for crop monitoring.
• The proposed system is suitable for practical application in diverse agricultural data analyses and research.