A robotic complex for automatic detection and removal of defective parts on a conveyor has been developed in Russia. The system was created by engineers from the Moscow Aviation Institute (MAI).
The developed robotic manipulator is designed to automate quality control on production lines. The complex identifies product defects directly on the conveyor and promptly removes defective items from the flow without operator intervention.
The key element of the system is a proprietary machine vision module. It processes camera images in real time, identifies parts with geometric deviations, surface damage, or marking errors, and sends a signal to stop the line.
After the conveyor stops, the MP-2 robotic manipulator takes over. It receives the coordinates of the defective part, precisely positions the gripper, extracts the item from the flow, and moves it to a storage unit for further analysis. Then the system resumes the line's operation. The entire cycle takes a few seconds.
The MP-2 robot has four degrees of freedom, can lift loads up to one kilogram, and positions the gripper with an accuracy of up to one millimeter. The machine vision system is based on OpenCV computer vision algorithms with the ability to retrain neural network models for specific defects.
According to MAI specialists, the implementation of this technology will reduce the share of manual labor in quality control operations and minimize the impact of the human factor.
