Minimization of Image Processing Resources and Deduplication for Recognition of Objects on Conveyor Belt

This study aimed to enable the recognition of the positions and rotation data of objects using the machine vision technology in the packaging process and the pick& place operations by robots using the data. For these objectives, this study proposed the methods of controlling the trigger signals of the camera and minimizing the imaging processing CPU resources in consideration of the recognition range of the campera, size of objects and speed of the conveyor belt. In addition, this study also tried to resolve the problem of an object being recognized as separate objects in the multiple frames of the camera.