Linuxcnc with Computer Vision …

Hi,
As I know there was no enough topics about this subject , but I’m working on idea that ,
We know that the deviation in z direction for plasma cutting can be solved by THC , floating head ..etc , but I’m working on a welding machine , where the deviation on x, and y will be more .
So I used python openCV , to measure this deviation , and I was able to get the difference dimensions in x,y .
But I was not able to update the G-code simultaneously .
Any idea to update the g –code …

No need to change gcode for that, have a look at external offsets, should be much easier to get to what you need.
linuxcnc.org/docs/2.8/html/motion/external-offsets.html
PlasmaC should be already using external offsets to give you a starting point.

tommylight wrote: No need to change gcode for that, have a loot at external offsets, should be much easier to get to what you need.
linuxcnc.org/docs/2.8/html/motion/external-offsets.html
PlasmaC should be already using external offsets to give you a starting point.

I swear I replied with essentially this same thing last night. I must have never clicked send then closed the browser without realizing...

Hi .. Tom

This is a good advise to start , but as I watched some video , all external offsets was controlled manually , as explained i will use it for a welding machine as a tracking sensor for the welding gap ... still the idea not clear for me how to integrate the feed back from openCV to external offset to effect the motion of x, y simultaneously .

never mind ... thank you

snowgoer540 wrote:

tommylight wrote: No need to change gcode for that, have a loot at external offsets, should be much easier to get to what you need.
linuxcnc.org/docs/2.8/html/motion/external-offsets.html
PlasmaC should be already using external offsets to give you a starting point.

I swear I replied with essentially this same thing last night. I must have never clicked send then closed the browser without realizing...

Did you make the same typo ???
loot...

Hamel2020, did you get it to work? If so how successful were you? I plan to do something similar using a THCAD10 as the welding sensor.

Hi …
Actually I’m still working on it … I was able to read the distance (gap-welding width) between the two plates as the photo , I used the external offset as Tommy advise , but the offset examples in axis is limited specially for sine oscillating motion damping in Z-direction , ( this will be helpful in case of plasma cutting ) but for welding need damping in XY-direction .
Still I need the convert the image reading into a voltage signal to indicate the max_lim and min_lim for the damping oscillating .
I used opencv and python

What kind of camera are you using?

this one