The challenge: Shiny door hinges
Box with shiny door hinges
Can Door hinges be gripped automatically? For a customer, we were asked to test small metal hinges that tend to snag. Two aspects of bin picking in particular were at issue:
- Can the shape and surface of the parts to be gripped be detected by machine vision systems?
- Can the robot grab them from a box without any problems?
This task becomes particularly difficult for machine vision systems if the object to be gripped not only has a complicated shape but also a shiny metallic surface. A moving, i.e. irregular, shape further complicates the recognition process.
To demonstrate the feasibility of such a challenge, we set up a study with shiny door hinges in two different sizes.
The test setup was carried out with three LumiScanX, a 3D light field sensor. The lenses were set to 4 mm, 6 mm and with 8 mm. The sensors were each placed at different distances. The greatest distance was two meters. Afterwards, images of the door hinges were taken in both sizes.
Result of the test
Bin picking worked with all hinge types. The larger door hinges could be reliably picked even at a greater distance of approximately two meters. Since the smaller hinges require greater precision, the maximum working distance for these was approx. 1.80 meters.
Conclusion and next steps
Door hinge in the software interface
Our experiment ended after the successful object detection – the feasibility was proven, the customer was satisfied.
For the concrete implementation of a bin picking system for the door hinges, cooperation with the internal plant engineering department or an integrator would now follow. This partner would then select a robot with a suitable gripper that can optimally grip the door hinges. For example, a suction gripper would be possible: Since the door hinges are too heavy for a simple solution, this should suck in the hinge to be gripped at two points. Other optimized gripping solutions would also be conceivable.