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 the 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.

Tips for Working with Shapefusion

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

Tips for Working with Shapefusion

Postby eliscio » Mon Aug 31, 2009 3:41 am

Hello All,

I recently scaled up my calibration grid so now I can scan larger objects about 0.5m^2 (more or less). I did this torso and cranked up the smoothness settings. Overall, the individual scans were pretty good, but when I fused them together, I always get "bumps" or lines where the scans meet.

Do you have any tips on how to really get a clean, fused scan?

Attached is an image of a mannequin that I scanned earlier today and you'll see the lines on the body.

Thanks

Eugene
Attachments
Mannequin.jpg
Mannequin.jpg (8.13 KiB) Viewed 3646 times
www.ai2-3d.com
User avatar
eliscio
 
Posts: 230
Joined: Sat Mar 14, 2009 2:05 am

Postby hal » Mon Aug 31, 2009 9:00 am

Hello Eugene,

from my experience with Shapefusion, these junction lines and bumps appears only in two matched cases: when the borders of the scans aren't perfectly complanar (not a perfectly aligned) and with too much high value of Fusion parameter. In this case seems that ShapeFusion interpret the irregular borders of every single scans as a detail, so it preserve these errors like a real detail of the object.
If the alignment is perfect and there aren't gaps or crossing surfaces between a scan and others, with the right value of fusion, in relation of the scans trinagle density, all goes fine.

Do you have cleaned the borders of every single scans before to fuse? It can help...
Bye,
Mattia
User avatar
hal
Moderator
 
Posts: 1527
Joined: Thu Mar 01, 2007 4:08 pm
Location: Cavaria con Premezzo (VA), Italy.

Re: Tips for Working with Shapefusion

Postby eliscio » Mon Aug 31, 2009 1:45 pm

Salve Mattia,

Well, this is what I was thinking...if I go back and delete the borders of each individual scan, then they might fuse a bit better. When I tested this, I think it definitely helps. There are a combination of things I need to improve in my scan though because I am still using a red (cheap) laser. I actually pulled it out of the level and mounted it in a hand held flashlight. I think it works better this way.

So, the higher the Fuse parameter, the higher the level of detail?
www.ai2-3d.com
User avatar
eliscio
 
Posts: 230
Joined: Sat Mar 14, 2009 2:05 am

Re: Tips for Working with Shapefusion

Postby hal » Mon Aug 31, 2009 2:02 pm

eliscio wrote:So, the higher the Fuse parameter, the higher the level of detail?


Yes... or better: if your scans have, as e.g., 1 mm of precision, you can fuse with an high parameter of fusion to save this details.
The amount of details are stored into every single scan meshes, You can only choise how is the percentagge of details that you need into the final fusion. However, I'm still confused about the numerical value of the Poisson fusion. I don't know the exact relation between the number into Poisson field and the amount of triangles generated. I hope that Sven or Simon can help me to understood better. Usually I fuse all at 1024 (arbitrary value) for high definition meshes, 512 for medium and 128 for very low detailed meshes.
But I remeber that for the some action figure scans, I've fused at a bigger value.

Obviously there is an other important consideration about the details on every scans: if you have a fine detail everlapped with other scans that don't have the same quality for this detail, probably you smooth it when you fuse. In these cases I suggest to cut away all the "bad" parts of the scans that don't have the best survey of detail and after fuse all togheter (do this after alignment).
An other hint just to improve the sharpness of the edges and shape of a scan, is to duplicate the desidered portion of scan two or three times.
Leave they in its original position (overlapped the main scan) and fuse all. You can achieve a better quality for that particular "duplicated" area.

Ciao,
Mattia
User avatar
hal
Moderator
 
Posts: 1527
Joined: Thu Mar 01, 2007 4:08 pm
Location: Cavaria con Premezzo (VA), Italy.

Re: Tips for Working with Shapefusion

Postby Francois » Wed Mar 31, 2010 4:47 pm

hal wrote:
eliscio wrote:So, the higher the Fuse parameter, the higher the level of detail?


... However, I'm still confused about the numerical value of the Poisson fusion. I don't know the exact relation between the number into Poisson field and the amount of triangles generated. I hope that Sven or Simon can help me to understood better. ...


I desperately searched the forum for some kind of answer to this simple and old question and didn't see a single answer.

Sven ? Simon ? Can you please answer ?

Thank you !
Francois
 
Posts: 60
Joined: Sun Feb 14, 2010 6:09 pm

Re: Tips for Working with Shapefusion

Postby Simon » Mon Apr 05, 2010 2:11 pm

eliscio wrote:... I always get "bumps" or lines where the scans meet.

Hi Eugene,
as Mattia already wrote, these lines might appear due to inperfect aligned scans, which in turn might be caused by surface deformations. These deformations in turn might be caused by imprecise background calibration panels, which also might be the reason why you have to cranck up the smoothness setting.

Francois wrote:
hal wrote:
eliscio wrote:So, the higher the Fuse parameter, the higher the level of detail?


... However, I'm still confused about the numerical value of the Poisson fusion. I don't know the exact relation between the number into Poisson field and the amount of triangles generated. I hope that Sven or Simon can help me to understood better. ...


I desperately searched the forum for some kind of answer to this simple and old question and didn't see a single answer.

Sven ? Simon ? Can you please answer ?

This question is not easy to answer, because there is not a direct relation between number of triangles and Poisson Fusion resolution. The resolution parameter of Simple Fusion is the maximal length (number of voxel per side) of a 3D grid/3D volume in which the scans are splatted. After that the fused mesh is reconstructed from this volume using Marching Cube. In contrast to Simple Fusion, Poission Fusion does not use a regular grid, but an Octree. Here the Octree depth depends on the scan point density, which affects the number and size of triangles of the fused mesh.
In short: the resolution parameter can be regarded as the number of cells^3 of a 3D grid, but only if the density of the original scan points are dense enough to fill the cells...
Hope this helps a little bit. :D
User avatar
Simon
DAVID Developer
 
Posts: 423
Joined: Sun Jan 21, 2007 3:40 pm
Location: Braunschweig - Germany


Return to Shape Fusion

Who is online

Users browsing this forum: No registered users and 1 guest