Dear DAVID Community,

As you are probably aware, the entire DAVID company, team and product range has been integrated into HP Inc.
This forum is now read-only. Relevant content has been migrated to HP and merged into HP's Support Forums on November 1, 2016.

To start new discussions for 3D Scanning, please register and post your new topic at the HP Support Forums

Manual alignment

You can post your wishes for new features in future program versions here.

Manual alignment

Postby Godehardt » Tue Sep 25, 2012 8:11 am

Hello Simon and Sven,
in DAVID 2, it was possible to align two scans manually. After choosing the respective option in the align field, we only had to define pairs of points, one in the respective scans, and then the two scans had been aligned one to the other. In DAVID 3, I really miss this point since I have to align, for example, a straight plane (Ebene) to a scan of a plane which contains abrasions and drilled holes. This has to be done to measure the volumes of abrasions and drilling holes. When I try to register the two object files then they are not layed one upon the other but penetrate each other a little bit (maybe because the procedure tries to minimize the average of all deviations between the two scans). If you do not want to make this manual procedure available in DAVID 3, then you could introduce a generalisation of the free registration in that you allow to define in one of the two scans the region (or several regions) which should be used for the registration. THIS WOULD REALLY BE A GOOD OPTION!!!
Best regards
Erhard Godehardt
User avatar
Godehardt
 
Posts: 111
Joined: Tue Aug 05, 2008 12:08 pm
Location: Hellenthal (Germany)

Re: Manual alignment

Postby Khalid Khattak » Thu Sep 27, 2012 7:36 pm

Yes, The manual Alignment was a great feature. Please include
PC: Dell Inspiron i3, 2.27GHz,4GB RAM,64 bit windows 7
CAMERA:
David CCD Mono USB
LASER:
David 5mW Green line laser
PROJECTOR:
3D LED Mini Projector 250lumens, Native 1024x600
[b]http://free3dscans.blogspot.com/
Khalid Khattak
 
Posts: 942
Joined: Thu Apr 23, 2009 1:30 pm
Location: Pakistan

Re: Manual alignment

Postby MagWeb » Sat Sep 29, 2012 7:43 pm

Getting back ManualAlignment would be a good thing...
Much better would be the idea to restrict AutoAlignment to masked areas of interest.

Also, I´d love to see non rigid alignment
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Postby hal » Sat Sep 29, 2012 8:59 pm

I agree with Gunter. Non Rigid Alignment should be the booster and the new point of strenght of ShapeFusion.

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

Re: Manual alignment

Postby Godehardt » Sun Sep 30, 2012 8:35 am

Yes, Mattia and Gunter,
alignment by using restricted area only would make manual alignment nearly unnecessary. And this together with non.rigid alignment would open a completely new area of applications to DAVID: combining scans of broken objects (perhaps after reversing the normals in one of the scans so that the areas of both objects to be combined point into the same direction).
Erhard Godehardt
User avatar
Godehardt
 
Posts: 111
Joined: Tue Aug 05, 2008 12:08 pm
Location: Hellenthal (Germany)

Re: Manual alignment

Postby Sven » Mon Oct 01, 2012 2:05 pm

We agree too :D
We are planning a function that allows to "paint" the area of interest.
Won't be in 3.5 but hopefully soon.
Non-rigid alignment... yes that would be great but that's not so easy... :oops:
Thanks for your opinions!
Sven
User avatar
Sven
DAVID Developer
 
Posts: 1541
Joined: Mon Jan 22, 2007 8:59 am
Location: Braunschweig, Germany

Postby hal » Mon Oct 01, 2012 3:48 pm

... but we trust in your great skills, Sven! And we know what kind ok geniouses you are, Simon and Markus!
So... Surprise us. :D 8)

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

Re: Manual alignment

Postby illustrami.com » Thu Dec 13, 2012 12:48 am

Hello

I read the posts, I fully agree.

The 3 or more couple of points approach is pretty practical, fast, very useful just as the ol'good screwdriver, but also painting areas of interest or viceversa would be great

There are so many different circumstances, therefore I lean to say that the more methods are available the better it is.

Then I read about the "non rigid aligment", and wandering what can this be after googling about I saw the video .

To me looks much something alike he used the Z-brush approach for mesh deformation and implemented it in the aligning application.

For live subjects it is almost necessary, but also when due to bad calibration or for big and long objects in general there are severe deformations

In fact it is a very good algorithm.

I do it by hand when possible and necessary, I cut the interesting parts and begin to align it where needed aslo manually (you see here for example how the 3 couple of points can be helpful in many circumstances, even though the alignment might be not good because the scans ar deformed at least one can fast position pieces needed for futher elaborations) but it is really a long work of patience, better to have an algorithm to do it :D

I guess must be not easy but I also guess that it is a big piece of code if to be written from scratch (Again!!! :D , our world is plenty of code rewritten from scratch dozens of times :D ).

I Imagine that the alignment done painting areas of interest must be a very good start point in order to eventually reach then also non rigid alignment, I imagine that having at least some parts which can be aligned the traditional way then maybe is possible to begin bending the parts that are not alignable.

For example in the video this guy shows how he aligns two partial scans of himself once with the head leaning forward right and once leaning back, on top of this the shoulders are twisted.

The face is alignable, little part of the shoulders too, but the neck absolutelly not.

Therefore one can align the face and have a start position, then the pieces of the shoulders and have a end position, then comes the difficult part, an algorithm should find out out the best deformation of the neck in one partial scan in order to fit the neck in the other one.

In the video the process appears to be different from what I say here, first the shoulders looks alike to be twisted and aligned and then with several approaches the second scan is deformed to fit the first one. But we do not know what the algorithm does without showing it, therefore I would not be suprised if it approaches the problem in this way or others or does combines several kinds of approaches.

I think that in David itself and in Meshlab there is already much to dig out and learn from in order to have ideas on how to reach this, I also think that understanding the deformation algorithms of Z-Brush would help really much.

Bye
illustrami.com
 
Posts: 28
Joined: Sun Nov 18, 2012 10:58 pm


Return to Feature Wish List

Who is online

Users browsing this forum: No registered users and 1 guest

cron