Re: [hugin-ptx] Architectural plans alignment

90 views
Skip to first unread message

Carlos Eduardo G. Carvalho (Cartola)

unread,
Apr 8, 2013, 9:31:40 AM4/8/13
to hugi...@googlegroups.com

2013/4/8 Martin Isak Jansen <mar...@blomstereng.org>
I have a little experience with using hugin for panoramas, but this is way beyond my hugin experience. Any suggestions?

Have you already tried and stopped at any point? See that you haven't said exactly what is your hugin knowledge and kind of asked something that could be a huge tutorial... maybe you could try and pass your problems to us to make it easy to help you.

Don't use the wizard tab to do your project. I would use manual control points and optimize little by little, keeping one image without optimization at all.
Maybe first optimize y,p,r, then y,p,r,v, then y,p,r,v,b and finally maybe Everything without translation.

Cheers,

Frederic Da Vitoria

unread,
Apr 8, 2013, 10:15:58 AM4/8/13
to hugin-ptx
I believe there was a question about a similar problem a few months ago, but I can't find it again.


2013/4/8 Carlos Eduardo G. Carvalho (Cartola) <cart...@gmail.com>

--
--
You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group.
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugi...@googlegroups.com
To unsubscribe from this group, send email to hugin-ptx+...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
 
---
You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hugin-ptx+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 



--
Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » - http://www.april.org

JohnPW

unread,
Apr 8, 2013, 1:59:37 PM4/8/13
to hugi...@googlegroups.com
Hi,
It seems to me that any "wonky-ness" must be with the roof drawing. The plan, being a measured drawing and directly scanned should be nearly perfect. I have to admit, I don't follow how one would get a roof drawing from GIS, but I assume it was originally from an arial photo? Anyway, It seems to me, that the roof image should be stretched to fit the plan (which theoretically would be easiest to do manually. Set some identifiable key anchor points on the line drawing that correspond to points on the plan and stretch (perhaps the ridge-line or probable locations of the corners of the underlying building?)
But if this hasn't worked for you, pserhaps you should show us your result and the problems you encountered?

John 


On Monday, April 8, 2013 4:37:44 AM UTC-5, Martin Isak Jansen wrote:
Dear group!

Ingredients:
I have a bunch of architectural floor plans that is old (1918) and hand drawn. They are big scans.


I also have a new line drawing of the roofs that is in correct scale (from a GIS program).



What I want:
What I would like to achieve is to align all floor plans with the new roof line drawing so it all matches on top of each other and are in scale. I have tried manually stretch and rotate the drawings to match etch other, but the result is a bit wonky.
So I thought that this might be done better with some hugin magic. The resulting file could be separated files or one file with layes.

How:

JohnPW

unread,
Apr 8, 2013, 2:34:30 PM4/8/13
to hugi...@googlegroups.com
On looking at your example images carefully, they don't seems to be from the same building (the dormers do not appear to be arranged in the same way on the two plans, no matter how the images are oriented.) Do these two images go together, or are they just examples?
John

JohnPW

unread,
Apr 8, 2013, 4:21:53 PM4/8/13
to hugi...@googlegroups.com
This worked pretty well though:

plan.zip

Terry Duell

unread,
Apr 9, 2013, 1:07:21 AM4/9/13
to hugi...@googlegroups.com
On Tue, 09 Apr 2013 06:21:53 +1000, JohnPW <johnpw...@gmail.com> wrote:

> This worked pretty well though:
>
This seems to me to be a mosaic mode problem, i.e. similar to stitching
flat scanned images each with a different lens, and optimise y, Trx, Try,
Trz,.
As there is no real common detail for a sane stitch/blend, I guess it is
probably best to just output remapped images.

Cheers,
--
Regards,
Terry Duell

Terry Duell

unread,
Apr 9, 2013, 1:09:23 AM4/9/13
to hugi...@googlegroups.com
On Tue, 09 Apr 2013 15:07:21 +1000, Terry Duell <tdu...@iinet.net.au>
wrote:

> and optimise y, Trx, Try, Trz,.

Ooops, sorry, that was meant to be r,Trx, Try, Trz.

Terry Duell

unread,
Apr 9, 2013, 1:26:36 AM4/9/13
to hugi...@googlegroups.com
On Tue, 09 Apr 2013 15:09:23 +1000, Terry Duell <tdu...@iinet.net.au>
wrote:

> On Tue, 09 Apr 2013 15:07:21 +1000, Terry Duell <tdu...@iinet.net.au>
> wrote:
>
>> and optimise y, Trx, Try, Trz,.
>
> Ooops, sorry, that was meant to be r,Trx, Try, Trz.
>

Here is the .pto resulting from optimising r, Trx, Try, Trz, and using a
different lens for each image.
It optimised with a pretty small error, and the remapped images look like
they should line up OK, but I really haven't checked. One image needs to
be given some transparency and then overlaid on the other...not sure how
to that.
Plan - Roof.pto

Battle

unread,
Apr 9, 2013, 8:28:17 AM4/9/13
to hugi...@googlegroups.com
Martin,
Perhaps a display solution like this one would work.
Its in a password protected directory.  user=hugin p=guest.  
I was able to scale the two items in this case the photo and the original drawing to work together in the viewer.  
Battle

Battle

unread,
Apr 9, 2013, 12:05:53 PM4/9/13
to hugi...@googlegroups.com
Martin, Here are your images from the pto JohnPW posted in a comparison viewer.  
Battle

JohnPW

unread,
Apr 9, 2013, 2:48:34 PM4/9/13
to hugi...@googlegroups.com
That makes sense, Terry.
I need to think about the variables and only enable the relevant ones instead of over enabling them and assuming they will sort themselves out during optimization.
There are definitely problems with the roof drawing though. One is inaccuracy (from distortion in the base images used to make them?) and the other is that the two drawings don't match (dormers are different.) I think the basic idea should work though.

Terry Duell

unread,
Apr 9, 2013, 6:33:26 PM4/9/13
to hugi...@googlegroups.com
On Wed, 10 Apr 2013 04:48:34 +1000, JohnPW <johnpw...@gmail.com> wrote:


> There are definitely problems with the roof drawing though. One is
> inaccuracy (from distortion in the base images used to make them?) and
> the
> other is that the two drawings don't match (dormers are different.) I
> think
> the basic idea should work though.

I suspect that the scanned original drawings may not be as good as they
should. That image appears to have a number of creases and from experience
it can be difficult to fully flatten creases when scanning.

Carl von Einem

unread,
Apr 10, 2013, 6:56:43 AM4/10/13
to hugi...@googlegroups.com
Terry Duell schrieb am 10.04.13 00:33:
Also scans usually suffer from shear (g, t)
<http://wiki.panotools.org/Lens_correction_model#Image_shear_g_.26_t_parameters>

Carl
Reply all
Reply to author
Forward
0 new messages