Drone-Based Harvest Data Prediction Can Reduce On-Farm Food Loss and Improve Farmer Income

Area of Science:

• Agricultural Science
• Remote Sensing Technology
• Data Analytics in Agriculture

Background:

• On-farm food loss, particularly grade-out vegetables, presents a significant challenge in sustainable agriculture.
• Conventional methods for monitoring vegetable size and predicting optimal harvest dates are often cost-ineffective and labor-intensive.

Purpose of the Study:

• To develop and validate a comprehensive pipeline for automatic, non-destructive estimation and prediction of individual broccoli head sizes.
• To determine the optimal harvest date to maximize profit and minimize grade-out losses using a temperature-based growth model.

Main Methods:

• Utilized drone remote sensing and advanced image analysis techniques to estimate the size of over 3,000 individual broccoli heads.
• Integrated individual size data into a temperature-based growth model to predict optimal harvest timing.

Main Results:

• Achieved high accuracy in estimating and predicting broccoli head sizes across two years of field experiments.
• Demonstrated that deviating harvest by just 1-2 days from the optimal date significantly increases grade-out and reduces farmer profits.

Conclusions:

• The developed drone-based pipeline offers a cost-effective and highly accurate solution for crop optimization in broccoli cultivation.
• This approach provides a powerful tool for minimizing food losses and enhancing economic returns for farmers.