camera-supported robot guidance – cost-effective, precise and process reliable.
OCTUM's machine vision systems for robot guidance reliably perform the 2D and 3D position detection and workpiece identification in the production cycle. The results are transferred directly to the robot controller and thus allow cost-effective automation of typical robotics applications such as:
- unloading and loading of the workpieces
- palletizing and depalletizing
- sensor-aided part assembly
- identification, location and position determination of the workpieces etc.
The detection software works based on fast and simple trained contour models of the workpieces. This ensures the necessary robustness of the detection even in changing light conditions at the place of detection and variations in the size and position of the workpieces. This even works on workpieces which touch or partly overlap. The integrated calibration ensures that the results are transferred to the robot controller in world coordinates and can therefore be converted directly into motion.
2D robot guidance – proven, flexible robot vision.
If the part presentation plane is known, 2D detection systems usually suffice for the robot guidance with determination of the position of the part (x/y angle) in the known plane. Depending on the required resolution, accuracy and cycle time, both smart cameras and PC-based systems are used. The camera can be fixed stationary above the detection point or on the robot gripper. Combinations with a stationary camera e.g. for rough position determination or determination of the plane (by use of TOF cameras) with a hand-held camera is also possible. The machine vision system can of course also communicate with several robots simultaneously and contributes accordingly to inexpensive automation solutions.
3D robot guidance – when more degrees of freedom are required.
If the part to be picked up is tilted, a 3D detection system for robot guidance with determination of the part position with all degrees of freedom is used. At OCTUM this usually means an image processing solution based on stereo cameras, possibly combined here with TOF cameras for a rough position determination. The 3D "grip into the box" is not a goal for the use of our 3D systems for robot guidance.