IMAGE PROCESSING APPLICATION FOR ROBOTIC ARM TO RECOGNIZE AND PICK UP 3D OBJECTS MOVING ON CONVEYOR BELT

Abstract

The application of image processing to identify the position of moving objects on the conveyor belt and control the manipulator to pick up the object for assembly is one of the very popular applications in industry today. However, most of the applications are for 2D flat objects. 3D objects have not been studied much, especially in Vietnam due to the complexity of calculation and long execution time using traditional methods. This study presents a method that can increase the productivity of an assembly line by identifying and picking up 3D objects moving on the conveyor belt using image processing and manipulators. The system includes a camera placed above to detect the object to be assembled in any direction and position on the conveyor belt. Then the 3D coordinates and angular orientation of the object in the 3D coordinates are calculated. The calculated data will be transmitted to the manipulator to perform the picking task. The assembly process will be performed based on the known position and orientation of the object to check for errors. Through experimental verification, the accuracy as well as the recognition time, the system proves to be applicable to the assembly industrial environment to help increase the productivity of the system.