found it 
EDIT: had a look at the video, and the genius behind this is no other than @jamiek 
Such a little world
4 Likes
Dang i forgot about that!! That comes from Jaime. If he can dream it he can achieve it. I really dont think he sleeps!
2 Likes
Oh, I missed that. I thought you were trying to replicate the crib frame.
I have one. Very small. So first thing I did was make a much bigger frame, still trying to calibrate it for the sizes. The main issue I have with it is the web based vector creation and it leaves a LOT of remnants to clean up. It is fairly accurate, but I am not sure I will use it nearly as much as I want to want to.
Love the giant frame
I’m just wondering why they didn’t use the same tape as the shaper origin, but in reverse… Very odd…
So… It seems the weak point is still the vectorization process, not much better than other applications 
Maybe the smartest way is just attaching a camera on the core pointing down and using the LR3 as a giant flatbed scanner by stitching pictures
Possition data of the machine could help getting the size and stitching right…
Then you can use the image as a clean canvas, or try to vectorize it…
Some kind of gcode sender getting a webcam picture from a URL, moving the machine, and so on repeatidly… Then just some kind of scripting to stitch the pictures together…
I like this idea…
1 Like
I’m not sure if I’m missing something here, and I assume there’s a similar function in an Android Phone - I regularly “scan” odd shapes using my iPhone - I usually use photoshop to transform to the original shape, but you can use the crop feature in the phone to do a fairly good job.
Place the object on a square grid so that you have a reference.
Turn on “grid” on your phone and line it up as best you can with the square grid background. The phone has a pair of cross hairs (see the yellow and white crosses in the middle) that will align when it’s horizontal. Taking care to do that will eliminate most of the parallax error.
The crop tool has also three tools for rotation, horizontal and vertical parallax adjustments. Just fiddle until everything lines with the grid. It’s easier to do than to explain and I don’t have a grid to use in my hotel room! (I will do it with the bathroom tiles if this isn’t clear enough)
Then take the image and trace half of each side in your cad programme, and mirror each to get a perfect shape. Or just trace the lot if it’s not perfectly symmetrical to begin with.
For something like the job you are doing though, I would use a joggle stick (ticking stick) and whip it out with the bandsaw in less time that it takes to fire up the computer! 
EDIT - Couldn’t help myself - the same image before and after adjustment. Not perfect but this was an extreme example and I have a bus to catch!
5 Likes
Im insanely curious if you get this to work. I thought about it but I wasn’t sure how you would ever get all the points in the same place for the software to recognize it. My biggest complaint is the size of it.
And Ive been constantly wondering this. Seems like a $$$$ making decision.
It works, just hard to tell what the scaling actually is. For shapes it’s fine, but actual dimensioning it’s kind of hit or miss. I’m hoping shaper updates the software to be more user friendly, and increases accuracy.
So did you just measure the size of the circles and just scale it up? And then the app recognizes it? If so, I cant believe that worked. But it becomes much more useful if you can draw something that large.
What iPhone & iOS are you using? I don’t have the crosshairs.
Go to settings>camera and turn on “Grid” or in later models/OS turn on “Level” The cross hairs will appear when you hold the phone horizontally.
I have a 14 but this used to work on my 10 (X) - I keep both turned on as a matter of course - it’s great for copying documents too!
2 Likes
@bitingmidge , interesting. I regularly use the grid but the crosshairs never popped up. I guess I’ve never tried to take an overhead photo with the grid on. Thanks for the pointer!
1 Like
@jamiek Is this line follower feasible with the fluidnc based boards? I have the feeling that it has to be running on a pi or similar. Also is it something you use a lot?
I believe the frame of the shaper trace is for skew correction and scaling only, I don’t think it would be hard to make some software that does that and then outputs the transformed image to be processed later by Inkscape or something like that.
The line follower requires a Pi, in particular it’s a plugin for OctoPrint, so it’s not part of the controller. If OctoPrint were set up as a sender and sent commands to FluidNC then I guess that could work in theory.
I haven’t used it for any real work. It was purely a demo to see if I could do it.