Computer Vision

Hello dear Rotrics Dexarm users,
I have the following problem.
I ordered the Visions Kit “RB-Ror-03 Rotrics Computer Vision Kit” and received it 2 days ago.
I have mounted and connected the camera as described in the online instructions.
I am working with Chocolate Bean in my test, these are also recognized by the camera. And the Rotrics Dexarm also shuts down to grab them.
But now come the problems:

  1. the suction cup does not grip the Chocolate Bean, but it always descends a little before the Chocolate Beans and grips into the void.
    (A change of the values as indicated under point 3. start recognizing, shows no effect. The arm continues to grab at the same place!
  2. but when loosen the screws of the camera mount and I move a littel bit the camera offset, she grab one chocolate bean, but at the latest at the 2nd is the suction again next to it.
    Where or how can I set that in the software that the demo works!

Does anyone have an idea or has this possibly also someone tested and a solution for me!

Greeting Skink

Here attached pictures

  1. camera mounted at original position

  2. camera mounted slightly offset

Maybe you already did it, but a recalibration as explained here could work:

Just a suggestion, hope it helps

Hello DavidV,first of all thanks for the help.

I have of course already before and also after you have written a recalibration carried out, but without success.
However, I have read here in the description “DexArm movement accuracy problem” something else, where the problem could possibly still be.
And this text here : “2) The end-effector offset value
The Arm inverse kinematics also requires different offset values for different end-effectors. A wrong offset value will cause a large deviation in Y-axis.
Command M888 Pn can set the right offset value for different modules. Command M888 without P parameter can check the current module.”
Since the suction cup is suck 15mm before the chocolate lens and the offset value here is 105.00, I have tried to set the value to 120.00 with the command M888 Pn120.00, but the existing value of 105.00 does not change!
How must the parameter be used, maybe that’s where the error lies?

Hello Skink, it was obvious, but sometimes is worth to ask :slight_smile:
I found in this link there’s another way to recalibrate, I didn’t had a look yet:
But as you said it looks pretty clear it’s a “code” calibration saving the correct data. Hope someone from Rotrics might bring some light to the issue.