Dear DAVID Community,

In 2016, DAVID and this forum have been integrated into HP Inc.
This forum has been read-only since then, and will be switched off by then end of 2019.

To everyone who has contributed here, we thank you for all your work and for the constructive, helpful and friendly atmosphere you have created!
Please continue to discuss on the HP Support Forums.

Default orientation of fusion output

The place for all topics related to the 'Shapefusion' tool of DAVID.

Default orientation of fusion output

Postby potetz » Sun Mar 20, 2011 4:59 am

I am interested in relating the 3D coordinates of obj files created by fusion back to the image coordinates of my camera. I can do this for an obj file created from a single scan (by inverting the Tsai camera model stored in the config file). For fusion output, the only piece that is missing for me is the relation between the 3D coordinates of fusion output and the different 3D coordinates systems of each input scan.

My impression is that Shape Fusion chooses to express vertices within the world coordinate system of the final scan from the fusion process. In other words, if I fuse scan A to scan B, and then fuse the result to scan C, the output will use the same coordinate system as scan C. Is this true? By eyeball it seems true so far, but I can't tell if it will always be true (not just a fluke), or if it is exactly true. Can anyone verify this more concretely? I appreciate it.
Posts: 6
Joined: Mon Aug 02, 2010 8:39 am

Re: Default orientation of fusion output

Postby MagWeb » Sun Mar 20, 2011 8:34 am


hope I got you right.
If so: you're right:
If you align two scans (A + B) , and align A to B, the 3d coordinates of A are translated to the 3D world of B.
Now if you load a third scan C and align C to A , C's coordinates are translated to the world of A (which was registered in the world of B before).

To control in which scans world you are, you could load a saved scan twice. If you´d load scan B in the upper example again, B2 is exactly equal to B1.

(Note: Each fusion average nearby vertices - so you should do it one time only: When you aligned all scans.)

If you load that fusion result on the scanning page to the depthmap it is equal to the depthmap of scan B+shadows (lasershadows).

Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Default orientation of fusion output

Postby potetz » Sun Mar 20, 2011 4:14 pm

Thanks, Gunter. That's exactly what I needed to know.

I hadn't realized I could load an obj into David and then see its range image - that is pretty cool actually. For the result of a single scan this works as you say, but on my system (version 2.5.4), when I load the result of a shape fusion back into the "3D Laser Scanning" stage, the camera angle looks correct, but the range image is translated upwards a couple hundred pixels (partly out of the window). I don't think this will affect me, just fyi in case it indicates fishiness in the camera geometry handling.
Posts: 6
Joined: Mon Aug 02, 2010 8:39 am

Return to Shape Fusion

Who is online

Users browsing this forum: No registered users and 2 guests