Every single chicken fillet purchased in a store must first be harvested from a carcass. This is a task that is repetitive and tedious, so in order to increase production capacity and enable an increased use of raw material at an early stage in production, SINTEF has developed a robotic 3D vision system called the Gribbot that automates the harvesting process.
The "Gribbot" robot name derives from the resemblance of the gripper with the beak of a bird called a "Gribb." The Gribbot system features a robot arm from Denso Robotics, a Microsoft Kinect 2 camera, and a compliant gripper. The new Kinect model features an infrared 1080p wide-angle Time of Flight sensor— in comparison to the previous mode’s VGA resolution—that achieves a frame rate of 30 fps. It also features a monochrome CMOS sensor which captures video data and processes 2 Gbps of data to read its environment.
Programming for the robotic vision system was done in the LabVIEW graphical programming language and environment from National Instruments. In addition, the team developed a toolkit in LabVIEW for the Kinect 2, as well as made use of Denso’s robotics library.
“We all know it is difficult to automate tasks that rely on human hands. The challenge gets even better when one, in addition, has to consider the biological variation and the deformability of the raw material in contact with the gripper,” said Ekrem Misimi, Research Scientist (PhD) at SINTEF, Trondheim, Norway.
A rotating transport system is used to fix the carcass and present it to the Kinect camera. The 3D camera scans the carcass and localizes the gripping point—where the gripper should start the scrapping as part of the harvesting procedure of the chicken fillet from the carcass.
Page 1 | Page 2