Team members: Mohamed Ali A vision-guided robotic system is created to automate and optimize the packaging of white button mushrooms, addressing labor shortages in the mushroom industry. The system combines a high-speed delta robot, a vacuum suction end-effector, and a 3D vision-based imaging module to replace labor-intensive manual processes with precise, high-throughput weight adjustments. Using RGB and depth mapping, the imaging system segments individual mushrooms by size, orientation, and position, enabling the robot to perform precise pick-and-place operations that optimize till weights in real time. Advanced computer vision algorithms, including transformer-based instance segmentation, ensure accurate detection and gentle handling, while an offline weighted buffer queue dynamically updates mushroom weights for efficient adjustments. The system can handle up to 12 tills per minute and enhances efficiency, reduces waste, and maintains product quality, offering a scalable solution for other delicate crops or food products. Future work will expand its capabilities and focus on industrial deployment.
Mushroom Sorting Automation
Food and Agriculture