Can David do Automatic ََAlignment?

[3dxcan]

I was wondering if David can do auto alignment after each scan sequence. I've seen professional scanners that do auto alignment using "targets" (small pieces of paper stickers in shape of black rings) on the surface of the part or rotating table itself. I couldn't find the option in David 5.
Please help if auto alignment is possible in david 5.
target points:
http://proto3000.com/assets/uploads/Ima ... argets.jpg

targets on scan part: (way too many, tho)
http://proto3000.com/assets/uploads/Ima ... arkers.png

[3dxcan]

ok, No answers, so I answer myself, David 5 has removed the option to auto align to previous scan and it's got no option for aligning using target points. actually no one on this forum has ever seen or worked with target points.
Thanks everyone for helping out.

[avogra]

Hi 3dxcan,

here comes a late answer from me:
I'm not totally sure, if I have the same understanding of auto-alignment as you. In David, automatic alignment is the norm in the sense, that it will find the correct orientation between scans by itself. There is no extra option for auto-alignment. But you have different options to initiate this alignment.

• David 5 will align all scans of a sequence automatically after you finish the sequence. If I remember right, you can also start such an alignement sequence later, although this is a bit hidden. I will confirm this tonight for you.
• You can select two scans to to initiate an alignment of the the first to the second scan. This also works for grouped scans.
• If auto-alignment struggles, you can help it by marking matching points on each scan or by using the texture as extra help or by manually positioning scans

I'm not totally sure about photogrammetry, so what I say here is what I have read somewhere:
Those target points are more suitable for photogrammetry, where 3d-information is generated from the texture. 3d-scanners based on structured light like David3d usually don't need or use the texture for generating 3d-information. In contrast, purely white objects are the easiest to scan. In David3D you have to explicitly enable texture grabbing. There are structured light scanners where you have to buy an extra addon-camera to get texture at all, so target points don't make sense.
David3D uses the form of the object to align scans to each other. So an equivalent to the target points would be characteristic shapes you place on the object or the turntable. I like to use those wooden figures used in board games like this: Figurenkegel If you place them on the object, you have to remove them in post processing.
You also have the option to apply target points and enable "Use texture" in David3D.

[avogra]

I just read some more about those target points. It seems like many photogrammetry scanners need them to align all the photographs taken from different viewpoints to each other before reconstruction of the 3d surface. So you have to make sure, that the scanner knows how the pictures belong together and in the end, you have one complete 3d model.
Structured light is working quite differently. Here you do a single scan from one perspective and you already get one 3d surface. Of course every surface that was hidden is missing. That is why you have to take several scans and in the end align and merge those 3d surfaces into a closed 3d model.

[Micr0]

ok, No answers, so I answer myself, David 5 has removed the option to auto align to previous scan and it's got no option for aligning using target points. actually no one on this forum has ever seen or worked with target points.
Thanks everyone for helping out.

Yeah. I was really upset when they removed this feature. Having it do an alignment after each scan gave the subject good time to settle from rotation before the next scan. I miss that and turn table control.

[3dxcan]

I just read some more about those target points. It seems like many photogrammetry scanners need them to align all the photographs taken from different viewpoints to each other before reconstruction of the 3d surface. So you have to make sure, that the scanner knows how the pictures belong together and in the end, you have one complete 3d model.
Structured light is working quite differently. Here you do a single scan from one perspective and you already get one 3d surface. Of course every surface that was hidden is missing. That is why you have to take several scans and in the end align and merge those 3d surfaces into a closed 3d model.

Actually some of the higher end structured light scanners use those type of target points. GOM ATOS and Solutionix Rexcan are two of the high end SLS (no photogrametry) that I've seen align sequential scans using target points.
Please see this video: https://www.youtube.com/watch?v=r6cRBBcizMs

[avogra]

I see. And it looks like it is not the same as David using texture for alignment. Atos and Rexcan seem to explicitly detect those targets and are able to cleanly remove them from the scan in postprocessing.

[avogra]

of course I forgot to look up, if/how you can manually start alignment of a sequence. There is the option "Align sequence" if you right click into your list of scans. I think, this is exactly the same as the automatic alignment after you have finished a scan seqence. No idea why they didn't put this into the alignment window

[Mongrel]

Hi All,

David 5 does do automatic alignment after each scan, but with the David Turntable.

Mongrel.

[Thomas Sc]

Thanks, for the Information.
But most of us have self made turntable.
How we can do automatic alignment without com port?

Thomas