Records office preparation

[paul womack]

I am about to visit a record office. Their rules
permit cameras, but not tripods (let alone pano heads!)

I wish to capture the image of some 18th c maps, which are large,
in good detail. The obvious strategy is to take multiple
shots and stitch, but the shots will all be taken from different
position and angles (since they'll be taken freehand).

But since the maps are 2D a stitch should still be possible.

I have tested this approach at home (cheating using a tripod!)
and a road atlas.

I took two photos at (fairly extreme) positions and angles
as test/proof of concept. To ensure worst case
testing the two shots were taken at different distances, and with
different zooms.

I have hand set 14 control points between these two images,
and set horiz and vert controls on 1 image.

I am simply failing to get a high quality stable stitch.

I thought the correct approach would be to "fix" the lens for
one image, and optimise for YPR *and* XYZ for both images.

This has simply failed.

I would welcome advice.

BugBear

[Bruno Postle]

On 6 May 2014 09:20, paul womack <pwo...@papermule.co.uk> wrote:
> I am about to visit a record office. Their rules
> permit cameras, but not tripods (let alone pano heads!)
>
> I wish to capture the image of some 18th c maps, which are large,
> in good detail. The obvious strategy is to take multiple
> shots and stitch, but the shots will all be taken from different
> position and angles (since they'll be taken freehand).
>
> But since the maps are 2D a stitch should still be possible.

This can be done, I did the same with a painting that I wanted in
detail (though I had to the additional problem of having to wait for
the gallery attendant to leave the room).

The main issue is shooting handheld with insufficient or badly angled
light, see if they will let you use a flash. The other problem was
that surfaces that look flat are probably not, so try and have the
camera as perpendicular as possible, and maybe take some weights to
pull out the corners of the map.

In terms of aligning in Hugin, you can get this to work, but mosaic
optimisation isn't as stable as normal panorama alignment; the trick
is to get everything nearly right with a small number of photos and as
few parameters as possible, then gradually add photos and optimisation
parameters. A good tip would be to shoot a single overall shot and use
this as an anchor image to align the detail photos, then delete the
anchor photo once you have a rough alignment.

--
Bruno

[paul womack]

Bruno Postle wrote:

>
> In terms of aligning in Hugin, you can get this to work, but mosaic
> optimisation isn't as stable as normal panorama alignment; the trick
> is to get everything nearly right with a small number of photos and as
> few parameters as possible, then gradually add photos and optimisation
> parameters. A good tip would be to shoot a single overall shot and use
> this as an anchor image to align the detail photos, then delete the
> anchor photo once you have a rough alignment.

I have notably failed to get a good match
with the TWO pictures in the PTO I posted. That's
the smallest number possible.

BugBear

[Emad ud din Bhatt]

Make sure to minimize camera shake. Set camera on burst continuous shooting mode and take each frame multiple times. Your chance to get one sharp image will be maximized. Another trick is to take a table stand with you. Fix camera on tabletop and press tabletop+dslr against your chest to get sharp images handheld.

--
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
--- 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+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/5368BD83.5060101%40papermule.co.uk.

For more options, visit https://groups.google.com/d/optout.

--


Emaad
www.flickr.com/emaad

[paul womack]

Emad ud din Bhatt wrote:
> Make sure to minimize camera shake. Set camera on burst continuous shooting mode and take each frame multiple times. Your chance to get one sharp image will be maximized. Another trick is to take a table stand with you. Fix camera on tabletop and press tabletop+dslr against your chest to get sharp images handheld.

Can anyone suggest an optimisation setting
that will get the measly two images in my PTO
file to align (more or less) perfectly?

BugBear

[paul womack]

I have now tried leaving all the parameters
on my image 0 zeroed, and only optimising
the 6 positions parameters (YPR, XYZ)
on image 1. The optimisation is stable,
but the CP errors are dreadful;

average control point distance: 14.166496
standard deviation: 13.702146
maximum: 49.047939

There is clearly something about the Hugin/Panotools
model that I do not understand.

BugBear

[Carlos Eduardo G. Carvalho (Cartola)]

Just a fast consideration (I'm in a little rush) you could consider using a philopod next time. I've been using recently, even having considered it useless in the past, and I am pretty much satisfied now.

https://www.youtube.com/watch?v=ouOEM4cKKGc

http://www.philohome.com/tripod/shooting.htm

The video shows a very good approach to use 4 images in a full frame camera. Here an example I did and another one.

Bests,

Carlos E G Carvalho (Cartola)
http://cartola.org/360
http://www.panoforum.com.br/

--
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
--- 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+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/53689B5B.80309%40papermule.co.uk.

[paul womack]

Carlos Eduardo G. Carvalho (Cartola) wrote:
> Just a fast consideration (I'm in a little rush) you could consider using a philopod next time. I've been using recently, even having considered it useless in the past, and I am pretty much satisfied now.
>
> https://www.youtube.com/watch?v=ouOEM4cKKGc
>
> http://www.philohome.com/tripod/shooting.htm
>

> The video shows a very good approach to use 4 images in a full frame camera. Here an example I did <http://cartola.org/360/2014/04/27/tour-virtual-no-flisol-2014-em-niter%c3%b3i/> and another one <http://cartola.org/360/2014/04/22/feira-do-livro-na-escola-em-tour-virtual-360%c2%ba/>.

That's a great way to shoot a panorama, but that's not what I'm doing.

I'm trying to capture a large map spread on a desk or table.

I'm pretty happy with what my shooting technique HAS to be, since the records
office offers me little freedom.

I still cannot optimise my 2 shot example, and I simply do not understand
why. The 6 parameter model should (surely) have enough degrees
of freedom to represent my shooting conditions.

BugBear

[T. Modes]

First, you are using horizontal and vertical control points, so set output projection to rectilinear, otherwise this does not work correctly.

Second, I had some success by optimizing ypr from image 0 and ypr and TrXYZ+Tpy+Tpp for image 1 (you will need version 2014.0 for this). I got an average error of 4 and a max error of 15. Without the seeing the images, it is difficult to say, if this is good.

[David W. Jones]

On 05/05/2014 10:20 PM, paul womack wrote:
> I am about to visit a record office. Their rules
> permit cameras, but not tripods (let alone pano heads!)

Hmm, not even a monopod?

--
David W. Jones
gnome...@gmail.com
wandering the landscape of god
http://dancingtreefrog.com

[paul womack]

David W. Jones wrote:
> On 05/05/2014 10:20 PM, paul womack wrote:
>> I am about to visit a record office. Their rules
>> permit cameras, but not tripods (let alone pano heads!)
>
> Hmm, not even a monopod?

(chuckle) I've been looking into just how far
I can stretch their photography rules,
but I'm also looking into how good a job
I can do within their rules.

Further, record offices vary. One of them
basically says:

"No flash, since it can cause damage,
but otherwise, as long as you don't damage the
artefacts or impede other people, use what you like"

Another one says "no flash, no supports" (supports is an irksomely inclusive term).

BugBear

[paul womack]

Thank you.

I will try to duplicate your result. My control points are probably within 2-3 pixels, so I see no
(mathematical) reason not to hope for a result with a maximum error near to that.

BugBear

[bugbear]

Battle wrote:
> My experience with this kind of thing is to maximize the distance to the object using a longer lens when possible which reduces the angles and amount of variation from image to image which in the end tends to overcome the variation of handheld image capture. For example, can you put the maps on a floor, and stand on a chair? That would increase the distance from camera to object from 2 feet to 8 feet. Is there a balcony in the room that you could shoot down to a table on the lower level?
> Now we're talking 15 feet or more.

Yes - back in 2014 (which is when this thread is from) I worked the following out:

https://groups.google.com/forum/#!topic/hugin-ptx/nynP68FT5WA

BugBear

> Would they let you put the camera on a monopole and hold it above your head to shoot down to the table? That might get you 4 to 6 feet away. Even if all this were not permissible, maybe you could shoot tethered, and just hold the camera above your head looking at the computer screen to frame the shot. If you can shoot the map in two images, then you can probably hold the camera in place close enough to use the bracket feature to get 3-9 shots from each
> position with one shutter press. Chances of getting usable images when shooting handheld this way increase as others have pointed out.
>
> Also any relief change (from a flat plane) will lead to parallax issues. Getting the maps to lay flat as possible is more critical the closer you are to the maps. Do they have a display (drafting) case or table, or other table large that has a tilting top which will tilt 45 degrees? This would allow you to stand back from the maps far enough to shoot handheld.

>
> On Tuesday, May 6, 2014 at 4:20:43 AM UTC-4, bugbear wrote:
>
> I am about to visit a record office. Their rules
> permit cameras, but not tripods (let alone pano heads!)
>
> I wish to capture the image of some 18th c maps, which are large,
> in good detail. The obvious strategy is to take multiple
> shots and stitch, but the shots will all be taken from different
> position and angles (since they'll be taken freehand).
>
> But since the maps are 2D a stitch should still be possible.
>
> I have tested this approach at home (cheating using a tripod!)
> and a road atlas.

> <snip>

> I would welcome advice.
>
> BugBear
>

> --
> A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
> ---
> 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 <mailto:hugin-ptx+...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/19de5c19-a669-43f1-b114-cc343845f6c1%40googlegroups.com <https://groups.google.com/d/msgid/hugin-ptx/19de5c19-a669-43f1-b114-cc343845f6c1%40googlegroups.com?utm_medium=email&utm_source=footer>.