up looking camera advanced camera calibration leaves me stranded

52 views
Skip to first unread message

NismoBoy34

unread,
Jan 28, 2022, 7:26:20 AM1/28/22
to OpenPnP
Hello all hope you are well . 
Thank you for the advice so far its gotten me to progress .
Ive had a few issues with possibly a corrupt machine.xml ( reference head H1 would not open reference Nozzle ) I ended up deleting and recovering from a previous machine.xml .

Now in doing that ive noticed that bottom camera placements ( with adjustment ) are far from the correct co ordinates on the PCB .
If i disable bottom vision correction i get semi perfect placements for the 0603 components the few i have tested.
Please kindly help . Possibly i have missed something .
Thank you in advance .


machine.xml
vision-settings.xml
uplooking camera after advaced calibration .jpg

Jan

unread,
Jan 28, 2022, 8:40:09 AM1/28/22
to ope...@googlegroups.com
Hi Nismo!
This sound like a vision pipeline setup issue. Did you checked the
visual alignment process?
- put a part on the nozzle and move it over the bottom camera
- use the "Test Alignment" button on the Bottom Vision Settings tab of
the part.
- if the result is not satisfying, adjust the pipeline using the button
on the same tab.

Jan

PS: I've done the same procedure yesterday and found, that the
"Threshold" needs adjustment.
> --
> 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
> <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/266e0988-0383-44e0-82a1-5aef88e02f3en%40googlegroups.com
> <https://groups.google.com/d/msgid/openpnp/266e0988-0383-44e0-82a1-5aef88e02f3en%40googlegroups.com?utm_medium=email&utm_source=footer>.

mark maker

unread,
Jan 28, 2022, 8:51:45 AM1/28/22
to ope...@googlegroups.com

There instructions and a small animation in the Wiki:

https://github.com/openpnp/openpnp/wiki/Bottom-Vision#tips

NismoBoy34

unread,
Jan 28, 2022, 8:56:38 AM1/28/22
to OpenPnP
He @mark confirm is the picture exclusion correct from what you can tell in the picture above ?
PS i have installed a green card on my nozzles to help and assist with part identification .
Will run through the wiki again but seems i have followed the steps correctly a few times ( have run calibration about 15 times ) but incase i missed something will give it a once over again.

NismoBoy34

unread,
Jan 28, 2022, 8:59:45 AM1/28/22
to OpenPnP
If i run a test alignment with out move ticked i get  a reasonably good result ...
Then with move ticked the nozzle moves away from the center after checking the part as per the second image .
Part CEnter detect movement after ticking move to center.jpg
test alignment with out move to center .jpg

mark maker

unread,
Jan 28, 2022, 9:00:38 AM1/28/22
to ope...@googlegroups.com

I don't understand your question and/or what that strange image in the screen-shot is.

_Mark

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/562a308c-9d32-4339-9fe5-6bce98595210n%40googlegroups.com.

NismoBoy34

unread,
Jan 28, 2022, 9:02:47 AM1/28/22
to OpenPnP
Also have run a bottom vision pipeline to be certain .
Rect drawn with orientation .jpg

NismoBoy34

unread,
Jan 28, 2022, 9:03:59 AM1/28/22
to OpenPnP
@mark this image attached seems to block out the important part of the Nozzle and the part from view 
uplooking camera after advaced calibration .jpg

tonyl...@gmail.com

unread,
Jan 28, 2022, 9:25:49 AM1/28/22
to OpenPnP
>this image attached seems to block out the important part of the Nozzle and the part from view
Is that image taken with Advanced Camera Calibration enabled and the others with it disabled? That image appears to be from a camera that is rotated and tilted too much so that the vertical pixel (where the crosshairs intersect) doesn't even fall within the valid pixels - the rotated rectangular part in the upper left are the valid pixels and the crescent shaped part appears to be an artifact of lens distortion correction.  Did you physically adjust your camera and/or re-run the calibration between some of the images you posted? I looked at your machine.xml and it does show you have quite a few outlier points so I suspect you may also have some problem with properly detecting the nozzle tip during the calibration process.

Tony

NismoBoy34

unread,
Jan 28, 2022, 9:56:54 AM1/28/22
to OpenPnP
@tony thank you for the reply as it would seem i had that thought and made a crude but stable form of up camera calibration using a mirror . ive aligned as best i can using the camera sensor as the focal central point . Surely this would give good enough results from a parallel vision perspective ? If im wrong please advise where i could run correction ?
bolttom up camera central parallel .jpg

NismoBoy34

unread,
Jan 28, 2022, 11:18:25 AM1/28/22
to OpenPnP
@ Jan thank you for the tip we are on the road to success ! Thank you @tony and @Mark
Please see example of placements didnt have double edge tape so used some Flux to keep it sticky 
placement .jpg

tonyl...@gmail.com

unread,
Jan 28, 2022, 1:02:15 PM1/28/22
to OpenPnP
>  If im wrong please advise where i could run correction ?

Either run it via Issues and Solutions or follow the instructions here.

Reply all
Reply to author
Forward
0 new messages