Small parts vision issues

[Marshall S. (Alakuu)]

Trying to tweak vision for specific small parts.
I'd like to circle mask down very small for small parts. 
I can't seem to modify that in the specialized vision pipeline at all.
It states "Controlled by Pipeline Caller: Diameter="
None of the sliders on the parts calibration page seem to exist.
I remember on nozzle setup tweaking the circular mask to help ensure vision isn't looking at the wrong circle. Why can't I change this anymore for specialized parts vision?

[bing luo]

https://youtu.be/cWXWFRQNI8M     In the 4 minutes and 15 seconds of this video, there is a solution to it. I don't know what caused it, only Mark knows why.

[bing luo]

Change the name of Maskcircle, and its value can be modified

在2024年7月13日星期六 UTC+8 10:55:58<alaku...@gmail.com> 写道：

[Marshall S. (Alakuu)]

Fantastic! I did figure out I think what happened. It's limited by the nozzle 'maximum part diameter.'
When you change that the circle mask value changes. But it can't be specialized. Thanks Bing for the solution!

[mark maker]

> Change the name of Maskcircle, and its value can be modified

Don't.

First, please explain what the actual problem is. Why do you need a tighter mask?

Then maybe read about Max. Pat Diameter:

https://github.com/openpnp/openpnp/wiki/Setup-and-Calibration_Nozzle-Setup#nozzle-tip-configuration

_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/2f8c50c7-c81f-4c34-9c55-711ac15c2c93n%40googlegroups.com.

[bing luo]

I have encountered this problem in the past, and then I searched for the same case on Google. The result of your discussion at that time was to solve it in this way. But now I can't find that case anymore.

[mark maker]

> None of the sliders on the parts calibration page seem to exist.

Proceed as described here:

https://github.com/openpnp/openpnp/wiki/Computer-Vision#make-stock-vision-the-default

Further info:

https://github.com/openpnp/openpnp/wiki/Bottom-Vision#using-the-stock-vision-settings

_Mark

I remember on nozzle setup tweaking the circular mask to help ensure vision isn't looking at the wrong circle. Why can't I change this anymore for specialized parts vision? --

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/b173965a-b7f1-439a-a705-85290c988673n%40googlegroups.com.

[Marshall S. (Alakuu)]

Sure! 

My 0402 nozzle has a bit of a vision issue where the edges have some scratches that vision detects as pads. My SOD 323 diodes are garbage and their feet really aren't very reflective too.
I've been putting black primer on the nozzle but it peels after a while. 

I had completely forgotten about the max part diameter setting. Once I found it I was golden. Just me being dumb.

Once I shrunk the max part size down it worked great. On the super small nozzle I'll never pick anything large up anyway and thankfully no more vision mismatches. 
I still get vision failures but I'm alright with having it just set to defer.

Side question, is there a way to have it on vision part size failure to discard the part and pick another for a retry?

[mark maker]

> Side question, is there a way to have it on vision part size failure to discard the part and pick another for a retry?

No, not currently. See this answer (specifically point 2):

Quoting> Regarding retry after vision failed: OpenPnP does not currently support this. It only supports retry in the feed&pick loop and in the alignment itself, i.e. it retries the bottom vision operation, which happens so fast you might not notice (check the log). A retry loop across JobProcessor steps is not supported, because everything is currently done "per nozzle-full". On multi-nozzle machines it would have to go back to the feed&pick step but then only for the failed nozzle(s). Such a "masked" operation is not currently supported. And single nozzle machines get no special treatment. Currently, as Jan already mentioned, the only advice is using vacuum sensing, as it can happen inside the feed&pick step, i.e. inside the retry loop.

https://groups.google.com/g/openpnp/c/iG9u0yFtt8w/m/jhhVhzTnAgAJ

A solution would have to discard the part, unallocate the placement from the nozzle, and re-schedule it in the JobProcessor, but otherwise continue with the other nozzles which already have parts picked. In a single nozzle machine, it would immediately go back to do a next feed&pick. Such a thing would require a global retry counter, otherwise it could result in whole tapes discarded, when there is a persistent underlying problem, such as the bottom camera being blocked by a part that fell on it, or changed ambient light spoiling vision.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/511ae600-3b1b-420c-b152-b47ce21b5fa1n%40googlegroups.com.

[Marshall S. (Alakuu)]

That's fine. Complicated stuff to say the least. I've found defer works great for these pesky diodes I just end up pressing the run button twice and if it's telling me all parts are placed I just stop the job again.

Thanks for all the help! Amazing work as always Mark.