Current Status

We have successfully delivered an integrated ecosystem built upon three foundational fronts:

• Successfully developed and evaluated an advanced deep learning framework (including YOLOv8 and Grounded-SAM) for robust capsicum fruit localization, growth stage classification, maturity determination, and precise 3D peduncle detection using RGB-D sensing.
  
  

• Successfully designed, simulated, and fabricated a precision-engineered 5-DoF robotic manipulator arm and a compliant harvesting end-effector featuring adaptive gripping and horticultural shear cutting mechanisms.
  
  

• Successfully innovated and integrated an autonomous mobile robotic cart equipped with a damage-mitigating storage system and DEM-simulated impact-resilient fruit handling.
  
  

Together, these completed research fronts have successfully converged to establish a fully functional, seamless, and intelligent greenhouse capsicum harvesting ecosystem.

• The graphical abstract of the work is shown below: