Do lightburn is capable to provide precise camera positioning? I assume it’s depend of camera quality, resolution. Rest is math and math is same everywhere. By precision I mean something like this:
I need to precise set laser unit over item and mark/cut it.
From my knowledlge, to achieve it, camera must be near laser module and software need to know distance between laser point and camera center. Of course after calibration (distance, angle, distortion etc).
Question is if lighburn is “ready” for this or I need to search other software?
If can then is possible to add in near updates custom camera overlay to make cross, lines etc on camera view? Like on provided video.
Do installing 12, 16, 18 or even 108 MPx have any futher inpact to precision than 5 or 8 MPx?
Looking solution to use lighburn or aim to any expensive stuff (hardware + dedicated sw).
True, but the math must warp a distorted bitmap into a rectangular field, so the result cannot have uniform resolution or accuracy over the entire field. In round numbers, the accuracy will be on the order of a millimeter for a platform on the order of half a meter, both of which seem much larger than your requirements, but suggest the location error will be around 2%.
A fixture holding the object (or many objects) at a known location, with the laser accurately aligned to that fixture, will produce much better results.
Yep I agree, but I assume situation where camera is mounted on laser module and distance to surface will be 5-10 cm. It will capture small area and dissortion will not be measurable (for non wide angle optics). It’s like microscope. We can’t see disortion on small area.
I can do experiment, but important question is if 16 MPx (or more) camera make any sense?
For 400mmx400mm, I use a 8MP document caméra fixed with a 3d printed lock, it works very well & precisly (I would say less than 0.5mm), but there is something missing or something I missed… in Lightburn.
As soon as you put an object with a certain thickness, what you see is no more perfectly aligned to the camera, the top of your object will not appear at its correct location, and the error will grow with thicker material (and will be worst if your camera is nearer and object not centered) due to higher impact of perspective if object is near the camera, after camera have been calibrated at a certain height it would require an option to tell the height of the object to cut/engrave, so the perspective correction could be adjusted to take in consideration this height difference from the calibration process.
More pixels wont help in any case…
Having done a lot of programming in image processing and also in the field of automatically detecting & removing perspective from image (like picture scanner do nowadays), I am sure it is doable. Also the calibration process could be done most of the time fully automatilcally if lens is know rather than asking to click on each four points (at least if the support is a bit clean).
For your special case that seems to be on small object, you can do :
calibration on a smaller area lowering the camera, with the camera at the center X/Y of what you want to work on.
print the contour of your object onto the bed (or a support) and set your object lying on this draw perfectly aligned (no need of camera for this one and is the one that will give you the most precision).
Thank You for response.
Your case (camera on top and large area) is clear for me and it need manual re-calibration or smarter way (hieght sensor) to compensate material thickness.
I was thinking same as You provided. Camera will be mounted on laser module, so change height of laser will change also camera height. Distance to object will be always same (if it’s close to flat).
Will try with 16MPx camera, because I think that it may increase precision, because calibration need to zoom picture. Picture will be less blurry.
Anyway, overlay picture to camera view will be very helpful to precise calibration. For now I see workaround to use windows camera app. It have cross overlay or write own sw with OpenCV library. It just will be nice to have it in one sw.
Yes, this functionality I want to achieve. Lightburn have (beta) calibration option for camera mounted on laser module, but without better calibration possibilities and cross/lines layout it’s not helpful for now. I tested it and had bad results after calibration, probably of cheap camera (1280x720) and inverted X/Y axis P90 M50.
Idea is to mount good camera, very stable on laser module, do burn any dot/cross and find offset (X,Y shift) in camera control view. Not ideal way, but should work without any calibration. But I may be wrong
ps. It’s used in UV printers for precise printing on small objects.
Here is example of precise positioning for UV printing. It’s explain that is possible and how precise it can be (limited to step motors step):
You can use a head mounted camera in conjunction with the Print and Cut feature to pick up on fiducials - though I haven’t done anything with the head mounted camera system personally as of yet.
I saw video about it and it use head mounted camera. It’s not precise enought for my needs. Check attached video in my previous post. It explain what kind of precision I mean. It’s possible, but need to complete all things: good picture quality, resolution, stable mount point etc. Will order 8-16MPx camera for futher experiments.
I have experimented with a laser head mounted camera. I used a very cheap inspection camera (usb) for that, fitted with a laser pointer holder and corrected for X and Y. As I recall, the result was very accurate. The field of view is relatively small with this (my) camera and therefore there is no need for a particularly high resolution either.
I think you should try with what you have first, it might be ok.
PS. The outermost glass disc from the camera could not bear to be so close to the nozzle while I was engraving glass and stone, the surface was ruined, even though I always use compressed air.
Definitely. (but not resolution, no need to increase it when there is only such a small field of view)
There is also no reason to mount so far down. Resolution with the cheap camera was still much better as a camera mounted 60cm above the bed and it will also be if it is mounted a few cm higher than before. (remember, the field of view is only a few cm) The disadvantage is that the nozzle head itself gets more in the way of the image. However, I have not tried mounting this camera at an oblique angle to the focus point and compensating for the error in the program. It could be a small research project for the weekend
I have spent the morning mounting my experimental camera on the laser head and trying to get the camera calibrated.
Mounting requires a little extra attention to place it without it crashing into the 2 lower threaded rod holders of the machine bed, in the outer positions.
In order to use LightBurn I had to install version 1.2.0, something prevents the (head mounted) camera from running under version 1.3.xx, LightBurn just exits or freezes.
The calibration works according to plan until the last step, when the final calibration of the 4 crosses must be confirmed. The laser just drives home…
I have no idea what causes this, a bit of a shame especially when all 4 crosses are so close to each other. Because I didn’t perform calibration I can’t update the preview either. (live view in the camera window works well enough)
But to give you an idea of the actual resolution of the old inspection camera, I am sending some pictures.
