[hugin-ptx] Re: new projections

[paul womack]

Yuval Levy wrote:
>> We could also use a branch for those features that are not totally tested
>> (the input projections).
> 
> branches are a great tool, and so are tags. IMO tags are the tool to use 
> to mark points in time when SVN is more robust and stable than 
> experimental. New features can and should be added with little testing 
> to trunk to encourage experimentation. A branch would have made sense 
> for this case during the months in which the code for Jim's recent 
> commit was piling up. But at some point, branches needs to get back to 
> trunk. The sooner the better.

For anyone using a source control system, I would strongly
recommend reading this:

http://www.cmcrossroads.com/bradapp/acme/branching/

It has some excellent (and source control system
agnostic) overview of various strategies.

   BugBear

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: New Projections - examples and thoughts

[Yuval Levy]

Hi Tom,

Tom Sharpless wrote:
> It would be very easy to add interactive graphics (for drawing
> boundary lines and so on) to Panini.  They don't need to be programmed
> in OpenGL (which is a bit cumbersome) because Qt provides 2D painting
> at no extra charge.   And it would not be too difficult to display one
> adjustable projection at a time, perhaps with the "out of zone" parts
> hidden behind an overlay.

I think that would be just right for the tool.


> But if you wanted to show nicely blended multiple adjustable
> projections that could get very difficult.

nah, the GUI's function is not to impress through slick effects, but 
just to offer a crude feedback to the artist in operating the tool. The 
quality of the current fast preview in Hugin (which shows some artefacts 
at the 360° seam of equirects, for example) is good enough. The last 
mile at this stage happens in the brain of the author. The final 
rendering will be nicely blended and slick. The author just needs an 
indication to imagine what the result from his input will be.

> you really want to be able to make nice long gradual transitions from
 > one projection to another (as Max Lyons is doing in PTAssembler 4.5).

honestly I have not thought of it in these terms, but rather the other 
way around: there are "natural seams" between projections, i.e. single 
rows or columns of pixels for which two projections yield the same 
result. In the Architectural projection the natural seam between Miller 
and Lambert is at the equator. The corners of the cube are natural seams 
between rectilinear projections.

All I'd like to do is to paint boundaries (initially just horizontal or 
vertical, maybe later a combination of the two, and further further down 
the road any boundary) and assign a projection to the area inside the 
boundary, with the back-end calculating at what HFOV/VFOV/height/width 
combination two contiguous projections produce a natural seam.

Going beyond that, of course it would be nice to add factors of all 
sorts to bend and warp the projections so that they *do* fit, but as you 
say...


> So to get a useful projection control frontend I think we need to aim
> low.

and I could not agree more. Initially I'd be happy with a couple of 
sliders moving vertical lines that are in fact corners of a flat wall, 
and the back-end to calculate the parameters for the rectilinear 
projection on that wall.


> The most interesting question is, what projections are needed for a
> "zone system" of perspective?  I think there is still a lot more we
> can learn about that from the Baroque view painters.

I think the start is the equirectangular. And initially we should forget 
about zenith and nadir and work as if on a cylinder. cut it in slices 
and for each piece determine the scale for the projection so that its 
left edge fits the right edge of the projection to the right. There will 
be some scaling to make the vertical axis fit. Once this work, add 
another projection. Maybe panini. The challenge is to keep the model 
flexible to support later additions.

Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: New Projections - examples and thoughts

[Tom Sharpless]

Hi Yuv

On Jan 2, 4:05 pm, Yuval Levy  wrote:
>
> How difficult is it to add the UI tools to Panini?
>
> Ideally Panini would let the user define the areas and their
> reprojections while giving them real-time feedback.

Well, I _have_ been thinking about that kind of thing since I started
on pvQt.  At first I though it could develop into the front end of an
image alignment tool (but that needs overlayed rotating spheres, which
I don't know how to do in OpenGL).  I have been keen to try doing
something about perspective, too, so with the advent of the Pannini
projection (I think it is fine to spell it that way) that seems to be
what I'll be doing in 2009.

It would be very easy to add interactive graphics (for drawing
boundary lines and so on) to Panini.  They don't need to be programmed
in OpenGL (which is a bit cumbersome) because Qt provides 2D painting
at no extra charge.   And it would not be too difficult to display one
adjustable projection at a time, perhaps with the "out of zone" parts
hidden behind an overlay.

But if you wanted to show nicely blended multiple adjustable
projections that could get very difficult.  Hard edged slabs might be
OK (though ugly) for a lot of architecture, but you really want to be
able to make nice long gradual transitions from one projection to
another (as Max Lyons is doing in PTAssembler 4.5).  A program to show
& control that could turn into a full CAD / animation surface design
sort of thing, which is well beyond my capabilities.  There are 3D
graphic design packages that might be adapted, but I expect it would
take several energetic young experts several years to produce
something easy to use.

So to get a useful projection control frontend I think we need to aim
low.

The most interesting question is, what projections are needed for a
"zone system" of perspective?  I think there is still a lot more we
can learn about that from the Baroque view painters.

Cheers, Tom

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: New Projections - examples and thoughts

[Yuval Levy]

Tom Sharpless wrote:
> Happy New Year, Yuv
> 
> On Dec 31 2008, 7:18 pm, Yuval Levy  wrote:
>> What if the user could *arbitrarily* set the vertical lines and split
>> the pano into zones; and then assign projections to each zone? what
>> would be the constraints to make the zones seamlessly contiguous like
>> the cubefaces? which projections could be applied / mixed with wich?
>> what would we need to make this happen?
> 
> That is my dream too: an interactive program to fit the image
> projection to the subject.

Thank, you Tom, and everybody else who is sharing thoughts about this.

I thoughts more about it, and about what you write, and I came to the 
conclusion that there are two parallel chains: a visualization chain and 
a rendering chain.

IMO they should be connected by a common file format describing the 
process, whether it is a stitching process, a blending process, or a 
reporjection process.

Ideally I would use Panini (which I admit I have not had the time to 
play with yet) to juggle in real time with the image; define boundaries 
and projections; see how it will look like. The same way I would use 
hugin to juggle in real time with the images to be stitched.

The result would be an XML file describing input images, areas in them, 
projections and parameters to be applied in those areas. So that it can 
be fed to a batch tool / glue that calls the different command line 
tools to produce very high resolution output (gigapixel).


> Which is what the vedutisti did.

How about calling your tool "Vedutista", then? A "Vedutista" is somebody 
that plays with perspectives just like you describe.

To me, Panini reminds me of my youth, namely of 


and I think the perspective should be called Pannini to reflect the 
correct spelling of the Pannini's name.


> We have the technology to do that, but not the UI to contol it.

yes, that's exactly it. And applying the technology currently requires a 
lot of manual intervention.


> Maybe some of us could make a New Years' resolution to do something
> about that.

How difficult is it to add the UI tools to Panini?

Ideally Panini would let the user define the areas and their 
reprojections while giving them real-time feedback. Then the user could 
save a description of the result in XML format. something like the 
following (quick and dirty):
















the XML format still requires a lot of thinking, but if we do it right, 
we can set the hub to glue together the different technologies 
(raw-processing, deghosting, fusing, hdr-merging, aligning, stitching, 
blending, tone mapping, scaling, etc.) as convenient single purpose 
command line tools that can be mixed and matched in manual processing, 
or controlled by a batch processor that interprets such XML files and 
execute the necessary sequence of commands. And on top of it a fast 
visual tool to set the process / the XML file up.


> OpenGL can do most of the image processing to support an interactive
> GUI, at hardware speeds.   To get high-resolution output you would
> also need a back-end rendering system, however it could be a batch job
> driven by parameters set visually in the GUI.  That process would be
> very like the stitching and blending we do now, the main differences
> being: 1) each "source image" is a piece of an existing image,
> typically an already stitched panorama; 2) each source image gets a
> separate custom warp (or as I would prefer to say, "reprojection").

sounds like we are thinking of the same, more or less. The idea with the 
XML file is to extend it beyond stitching and slicing. It could be 
flexible enough to do any of these, with Hugin being the GUI for 
stiching, Panini the GUI for slicing (the name is justified, after all 
:-) ) and maybe other GUI tools for deghosting, focus stacking and the 
many other possible functions.


> I think it is doable.

I think so too.

Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: New Projections - examples and thoughts

[Tom Sharpless]

Sorry, left out the link to the Panini painting; here it is:

http://www.wga.hu/frames-e.html?/html/p/pannini/interio

-- Tom


--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: New Projections - examples and thoughts

[Tom Sharpless]

Happy New Year, Yuv

On Dec 31 2008, 7:18 pm, Yuval Levy  wrote:
> Hi all,
>
> After making the marvellous work of Bruno, Tom, Daniel and Jim
> accessible through hugin SVN I played with it a little bit
>
> 

Thanks for this page, I think a lot of people will find it
interesting.

>
> The Architectural projection got me dreaming about ZONING. Wouldn't it
> be nice to *dynamically* define different areas of the pano to be
> rendered with a different projection *seamlessly* ?
>
> Daniel split the pano in two zones, horizontally at the equator,
> assigning different projection to each zone. It works, because at the
> equator the two projections yield the same result.
>
> How about vertical splits? we already do that, e.g. when we transform a
> sphere into a cube - the four sides around the horizon are each a
> rectilinear projection with 90° HFOV, split vertically at four randomly
> choosen lines equally distributed on the horizon.
>
> What if the user could *arbitrarily* set the vertical lines and split
> the pano into zones; and then assign projections to each zone? what
> would be the constraints to make the zones seamlessly contiguous like
> the cubefaces? which projections could be applied / mixed with wich?
> what would we need to make this happen?

That is my dream too: an interactive program to fit the image
projection to the subject.

Which is what the vedutisti did.  They weren't stuck with a fixed
formula, just did whatever made the picture look right.  Case in
point:  Panini's view of Santa Maria Maggiore () which I recently
analyzed by comparing it to the ground plan of the cathedral.  The
horizontal spacing of the columns is exactly proportional to the true
view angle, as in a cylindrical projection.  But their heights (and
all other dimensions of the room) recede along exactly straight lines
to a single vanishing point, which is what you would get in a linear
projection but not a cylindrical one; and above all, horizontal lines
are also straight, as they would be in a rectilinear view but
certainly not in a cylindrical or "Panini" projection.

Having laid out the column spacing, the artist was perfectly free to
draw the horizontal and diagonal straight lines, then fill in the
image details.   A computer graphics nerd would say that he built a 3D
model of the church and then texture mapped the details onto it.  We
have the technology to do that, but not the UI to contol it.

Maybe some of us could make a New Years' resolution to do something
about that.


>
> Going from the front-end to the back-end:
>
> * James' fast preview is the ideal playground for this. The result is
> displayed / manipulated in real time. The following elements could be added:
> - a "line" tool to add/remove/move lines perpendicular to the horizon,
> at least initially. the assumption may be relaxed later on.
> - we'd need to assign a projection to the area delimited by two lines.
> - what kind of projections can be used / are compatible / produce a
> seamless contiguity?
> - for rectilinear, the center of the projection may be variable as well,
> though for a start just the middle is ok.
>
> * under the hood
> - we'd need to calculate some parameters backward for each zone so that
> they fit. what are the constrains?
>
> * rendering is probably the easiest part, just render and stack ?

OpenGL can do most of the image processing to support an interactive
GUI, at hardware speeds.   To get high-resolution output you would
also need a back-end rendering system, however it could be a batch job
driven by parameters set visually in the GUI.  That process would be
very like the stitching and blending we do now, the main differences
being: 1) each "source image" is a piece of an existing image,
typically an already stitched panorama; 2) each source image gets a
separate custom warp (or as I would prefer to say, "reprojection").

I think it is doable.

Regards, Tom

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Daniel M German]

 Yuval Levy twisted the bytes to say:


 >> We could also use a branch for those features that are not totally tested
 >> (the input projections).

 Yuval> branches are a great tool, and so are tags. IMO tags are the tool to 
use 
 Yuval> to mark points in time when SVN is more robust and stable than 
 Yuval> experimental. New features can and should be added with little testing 
 Yuval> to trunk to encourage experimentation. A branch would have made sense 
 Yuval> for this case during the months in which the code for Jim's recent 
 Yuval> commit was piling up. But at some point, branches needs to get back to 
 Yuval> trunk. The sooner the better.

Ok, let us leave it as is. 

--dmg

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuval Levy]

Daniel German wrote:
> In my opinion it is better to commit one "task" at a time, so each
 > can be isolated and debugged independently of each other.

atomic commits are a good practice indeed and your advice is very 
valuable for future commits.


> Why don't we roll back your changes and introduce one change at a time?

for this specific case I am against a rollback. I've built everything in 
Ubuntu and what I have tested works well and I did not find any serious 
regression / broken things. I am looking for how to test the mirror 
input projection, but I am afraid it will have to wait to next week as I 
am away from the mirror itself.


> We could also use a branch for those features that are not totally tested
> (the input projections).

branches are a great tool, and so are tags. IMO tags are the tool to use 
to mark points in time when SVN is more robust and stable than 
experimental. New features can and should be added with little testing 
to trunk to encourage experimentation. A branch would have made sense 
for this case during the months in which the code for Jim's recent 
commit was piling up. But at some point, branches needs to get back to 
trunk. The sooner the better.

my 2 cents
Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuv]

Jim Watters wrote:
>   added mirror ... as input image formats

I just realize this. Are these the "donuts" that we see from mirrors
like http://www.0-360.com/ ?

how do we add this to hugin?

I had one of these mirrors on the shelve for the last three years. And
I have an adapter to put it in front of my camcorder...

Yuv
--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Daniel German]

hi Jim,

Please don't take this as a criticism.

I think committing this many additions of features in a single commit
is too risky,
and you have mentioned that they are not very well tested. In my
opinion it is better
to commit one "task" at a time, so each can be isolated and debugged
independently
of each other.

Why don't we roll back your changes and introduce one change at a time?
We could also use a branch for those features that are not totally tested
(the input projections).


--dmg

On Wed, Dec 31, 2008 at 9:07 AM, Yuval Levy  wrote:
>
> Jim Watters wrote:
>> Thanks.  I made the corrections.
>
> Thanks, all works fine.
>
> In Hugin (I will committ to SVN soon) all the new projections are
> available and render well. The options in the stitcher tab may require
> some fine tuning (e.g. the calculation of optimal size) and the fast
> preview is quirky with Panini (try dragging). With Orthographic and
> Equisolid it has difficulty at the extreme (360°).
>
> Is queryfeature.c documented?
>
> while waiting for hugin to build I looked at it and noticed that some
> projections are mentioned in "LensType" while others, particularly the
> new ones, are not. And then there are those commented out, whose
> numbering is not consistent with "PanoType".
>
> Do I understand correctly that LensType is the input projection and
> PanoType the output projection?
>
> And: queryfeature.c mentions the anti-aliasing filters. How difficult
> would it be to access them from hugin?
>
> Yuv
>
>
> >
>



-- 
--dmg

---
Daniel M. German
http://turingmachine.org

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuval Levy]

Jim Watters wrote:
> Thanks.  I made the corrections.

Thanks, all works fine.

In Hugin (I will committ to SVN soon) all the new projections are 
available and render well. The options in the stitcher tab may require 
some fine tuning (e.g. the calculation of optimal size) and the fast 
preview is quirky with Panini (try dragging). With Orthographic and 
Equisolid it has difficulty at the extreme (360°).

Is queryfeature.c documented?

while waiting for hugin to build I looked at it and noticed that some 
projections are mentioned in "LensType" while others, particularly the 
new ones, are not. And then there are those commented out, whose 
numbering is not consistent with "PanoType".

Do I understand correctly that LensType is the input projection and 
PanoType the output projection?

And: queryfeature.c mentions the anti-aliasing filters. How difficult 
would it be to access them from hugin?

Yuv


--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Jim Watters]

Yuval Levy wrote:
Jim Watters
wrote:
  
I have committed some changes to libpano13

  
  hugin now quits with

hugin: queryfeature.c:429: panoProjectionFeaturesQuery: Assertion `0' 
failed.
Aborted

I look at queryfeature.c before line 429 is the 
panoProjectionFeaturesQuery function and I don't 
see your new projections there.
  
  around line 400, after

 case PANO_FORMAT_FISHEYE_FF:

add:

 case PANO_FORMAT_ORTHOGRAPHIC:
 case PANO_FORMAT_EQUISOLID:

sorry, I have no SVN access to contribute this fix.

  
  and: around line 95, PanoType14 and Panotype15 should read 
Orthographic and Equisolid, not Architectural...

Yuv

Thanks.  I made the corrections. 

-- 
Jim Watters

Yahoo ID: j1vvy ymsgr:sendIM?j1vvy
jwatters @ photocreations . ca
http://photocreations.ca


--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
  To unsubscribe from this group, send email to hugin-ptx-unsubscr...@googlegroups.com
  For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuval Levy]

Yuval Levy wrote:
> Yuval Levy wrote:
>> Jim Watters wrote:
>>> I have committed some changes to libpano13
>> hugin now quits with
>>
>> hugin: queryfeature.c:429: panoProjectionFeaturesQuery: Assertion `0' 
>> failed.
>> Aborted
>>
>> I look at queryfeature.c before line 429 is the 
>> panoProjectionFeaturesQuery function and I don't see your new 
>> projections there.
>>
> 
> around line 400, after
> 
>  case PANO_FORMAT_FISHEYE_FF:
> 
> add:
> 
>  case PANO_FORMAT_ORTHOGRAPHIC:
>  case PANO_FORMAT_EQUISOLID:
> 
> sorry, I have no SVN access to contribute this fix.
> 

and: around line 95, PanoType14 and Panotype15 should read Orthographic 
and Equisolid, not Architectural...

Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuval Levy]

Yuval Levy wrote:
> Jim Watters wrote:
>> I have committed some changes to libpano13
> 
> hugin now quits with
> 
> hugin: queryfeature.c:429: panoProjectionFeaturesQuery: Assertion `0' 
> failed.
> Aborted
> 
> I look at queryfeature.c before line 429 is the 
> panoProjectionFeaturesQuery function and I don't see your new 
> projections there.
> 

around line 400, after

 case PANO_FORMAT_FISHEYE_FF:

add:

 case PANO_FORMAT_ORTHOGRAPHIC:
 case PANO_FORMAT_EQUISOLID:

sorry, I have no SVN access to contribute this fix.

Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Yuval Levy]

Jim Watters wrote:
> I have committed some changes to libpano13

hugin now quits with

hugin: queryfeature.c:429: panoProjectionFeaturesQuery: Assertion `0' 
failed.
Aborted

I look at queryfeature.c before line 429 is the 
panoProjectionFeaturesQuery function and I don't see your new 
projections there.

Yuv

--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: new projections

[Jim Watters]

I have committed some changes to libpano13 that
  added mirror, equisolid, orthographic, and stereographic as input 
image formats
  added or update equisolid, orthographic, stereographic as output 
panorama formats
  add some more defines to panorama.h to help image and pano formats
  updated comments and documentation to help adding more projections and 
using the new ones.

The equisolid and orthographic, and stereographic projections came from 
the work Helmut Dersch was doing with the Motion Panoramas 
http://webuser.hs-furtwangen.de/~dersch/ project he was working on.

I did NOT manage to convert the equations without error into a format 
for libpano13. 
Panotools can create stereographic, but I don't have test images of the 
other formats to verify correct implementation.

I have been holding onto this code for almost a year.  It seamed like 
the right time to commit it.  With this resent changes there will be 
someone familiar with the code to spot the problem and fix.  I hope. :)

Equisolid, orthographic, and stereographic are all kinds of fisheye 
projection.  Some fisheye lenes optics are better suited using one of 
these projections.
Mirror should help creating panos from mirror balls 
http://wiki.panotools.org/Spherical_mirror_panorama

-- 
Jim Watters

Yahoo ID: j1vvy ymsgr:sendIM?j1vvy
jwatters @ photocreations . ca
http://photocreations.ca


--~--~-~--~~~---~--~~
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 hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---