Bulk Tea Shoot Detection and Profiling Method for Tea Plucking Machines Using an RGB-D Camera.

Area of Science:

• Agricultural Engineering
• Robotics
• Computer Vision

Background:

• Addressing the urgent need for mechanized tea plucking due to rural labor shortages and an aging population.
• Highlighting the limitations of existing tea plucking machines in hilly terrains, particularly in cutter profiling performance.

Purpose of the Study:

• To develop and validate an automatic cutter profiling method for bulk tea harvesting using RGB-D camera technology.
• To improve the efficiency and quality of mechanized tea plucking, especially in challenging environments.

Main Methods:

• Utilizing an RGB-D camera to capture depth information of tea shoots.
• Employing improved super-green features and the Otsu method for shoot detection and segmentation under varying lighting.
• Generating cutting poses based on depth values for precise cutter profiling controlled by an upper computer system.

Main Results:

• Achieved an average bud and leaf integrity rate of 81.2%, leakage rate of 0.91%, and loss rate of 0.66%.
• Demonstrated a qualified rate of 90.4% for harvested bulk tea shoots.
• Confirmed improved cutting pose calculation accuracy and adherence to machine plucking quality standards.

Conclusions:

• The proposed automatic cutter profiling method effectively addresses challenges in mechanized tea plucking.
• The system enhances harvesting efficiency and ensures harvested tea meets quality standards for processing.
• This technology offers a viable solution for modernizing tea plantations facing labor constraints.