Question about primary fiducial

[Jim Young]

I'm reading through the documentation about calibrating the primary fiducial and I can't quite figure out if the testing jig is a permanent fixture on my table or not. I hope not as I already have established part holders mounted. 

Also the instructions say to use only a printout of the fiducial hole PDF and nothing else. Why is that? A sprocket hole from a black tape is a much cleaner hole than anything my printers can produce.

[Ian Arkver]

"Alternatively you can cut a fiducial out of a spare PCB. Other users have cleverly used the sprocket hole from a black plastic carrier tape on bright ground (1.5mm diameter)."

https://github.com/openpnp/openpnp/wiki/Visual-Homing

Personally I use a bit of an old PCB.

Since Visual Homing is done every time the machine is homed the fiducial needs to be permanently fixed and must not move.

It can be anywhere on the table, so long as it's in view of the camera and fixed permanently.

Regards,
Ian

[Jim Young]

Yes, for the homing fiducial. I understand that is permanent. My question was about calibrating the primary fiducial.

Vision Solutions · openpnp/openpnp Wiki (github.com)

[Ian Arkver]

Ah, ok, that wasn't clear from your question.

I used the printout as suggested. That's what the vision is geared up for.  Once you've done those calibration steps you shouldn't ever need to redo them (in theory), so the fixture can be removed.

I stuck mine temporarily down to a PCB in the PCB holder.

Regards,
Ian

On Thursday, March 24, 2022 at 8:07:46 PM UTC infinite...@gmail.com wrote:

[Jim Young]

I ended up using a sprocket hole on paper that was painted black with a Sharpie (except for the hole, of course) . It was what openpnp would recognize. I tried printing the PDF on glossy photo paper and a laser printer, but the software would not recognize them at all.

[mark maker]

I have clarified the Wiki:

https://github.com/openpnp/openpnp/wiki/Vision-Solutions#calibration-primary-fiducial

Using sprocket holes in tapes is only admissible for very thin (usually black) plastic tapes, not for thick paper tapes. The latter have a 3D structure which basically appears as two ambiguous circular features when seen from the side. Note that the calibration needs to look at it from the side.

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/458d1056-c6ad-44dd-9afc-5d1ef1e0f979n%40googlegroups.com.

[Jim Young]

Thank you for your work!

Yes, I'm using the hole from a black plastic tape. I simply could not get the printed PDF to be recognized, no matter how fine a detail I tried to print it.

[mark maker]

> I simply could not get the printed PDF to be recognized, no matter how fine a detail I tried to print it.

I'm a bit surprised by that. The DetectCircularSymmetry method is usually very powerful to detect even very "bad" subjects.

Are you sure, you set the diameter right? Have you tried setting it manually using the up/down arrows ?

https://github.com/openpnp/openpnp/wiki/DetectCircularSymmetry#example-images-and-use-cases

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ed59361e-3afa-4155-b5aa-7afb6f500750n%40googlegroups.com.

[Jim Young]

Well, I tried manual setting the circle detection, but the green circle and cross never appeared. When I tried auto adjust it never seemed to fine the circular outline. I think it might have more to do with the quality of my print out, although I tried it with both glossy photo paper at high quality as well as a laser printer. There was a lot of 'noise' in the print out that I think was throwing the detection off.