Hi all,

I have a problem to generate w/ POV 3.5 an image of an existing 
landscape. I got DEM data of this area, converted them to PNG files but 
I have some problem to make the POV image like the reality.

The real view : http://mamaf.online.fr/Annecy/image13.htm

My image w/o angle option for the camera : 
http://destroyedlolo.homeunix.org:8080/~laurent/sans_angle.jpg

and the same one specifying angle 38: 
http://destroyedlolo.homeunix.org:8080/~laurent/angle_38.jpg

Has you can see, it seems far mountains appear flattened ...

Do you know a way to make my images more realistics ?

Thanks

Laurent

Post a reply to this message

destroyedlolo wrote:
> Has you can see, it seems far mountains appear flattened ...
> 
> Do you know a way to make my images more realistics ?

I guess your camara is too near. Try to increase the distance between 
the camera and the mountains and lower the camera angle to keep roughly 
the same display window.
This should make the near mountains smaller compared to the far ones.

Hope this helps,
Henry

Post a reply to this message

if all else fails, scale 2*y or similar could help?


"destroyedlolo" <l_f### [at] yahoocom> wrote in message
news:411### [at] yahoocom...
> Hi all,
>
> I have a problem to generate w/ POV 3.5 an image of an existing
> landscape. I got DEM data of this area, converted them to PNG files but
> I have some problem to make the POV image like the reality.
>
> The real view : http://mamaf.online.fr/Annecy/image13.htm
>
> My image w/o angle option for the camera :
> http://destroyedlolo.homeunix.org:8080/~laurent/sans_angle.jpg
>
> and the same one specifying angle 38:
> http://destroyedlolo.homeunix.org:8080/~laurent/angle_38.jpg
>
> Has you can see, it seems far mountains appear flattened ...
>
> Do you know a way to make my images more realistics ?
>
> Thanks
>
> Laurent
>

Post a reply to this message

Doppelganger wrote:
> if all else fails, scale 2*y or similar could help?

I'm not so sure because It will also rise the 1st plan hill.
Anyway, I wander what is wrong there, why I have a so different result 
in POV compared to the reality ...

Bye

Laurent

Post a reply to this message

news:411### [at] yahoocom...
> I'm not so sure because It will also rise the 1st plan hill.
> Anyway, I wander what is wrong there, why I have a so different result
> in POV compared to the reality ...

First, it can be supposed that the camera position, angle and look_at are
different in the photo and in your POV-Ray scene.

Second, keep in mind that the maximum height of a POV-Ray height_field is
always, at its maximum, equal to 1. Also, its length and width are always 1.
This may not be the same as in the DEM specification. If the DEM files use
real-life dimensions, for instance, then the converted scene should be
scaled accordingly :
scale <DEM_size_x, DEM_size_y, DEM_size_z>
Of course, the camera should then be placed also accordingly.

G.

-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters

Post a reply to this message

"Gilles Tran" <gitran_nospam_@wanadoo.fr> wrote:

> news:411### [at] yahoocom...
> > I'm not so sure because It will also rise the 1st plan hill.
> > Anyway, I wander what is wrong there, why I have a so different result
> > in POV compared to the reality ...
>
> First, it can be supposed that the camera position, angle and look_at are
> different in the photo and in your POV-Ray scene.

As it's not my own photo, I'll take another one this afternoon (if the
weather is not too cloudy) and I will provide also information given by my
camera.

> Second, keep in mind that the maximum height of a POV-Ray height_field is
> always, at its maximum, equal to 1. Also, its length and width are always 1.
> This may not be the same as in the DEM specification. If the DEM files use
> real-life dimensions, for instance, then the converted scene should be
> scaled accordingly :
> scale <DEM_size_x, DEM_size_y, DEM_size_z>
> Of course, the camera should then be placed also accordingly.

I think it's OK also because I set :
- DEM_size_x and DEM_size_z accordingly to the size of my DEM file,
- DEM_size_y is set to 65535 as I use 16bits PNG file.

Anyway, I will provide also DEM datafile and script I use.

Bye

Laurent

Post a reply to this message

news:web.41207b10e6dfd373e5987fa60@news.povray.org...

> I think it's OK also because I set :
> - DEM_size_x and DEM_size_z accordingly to the size of my DEM file,
> - DEM_size_y is set to 65535 as I use 16bits PNG file.

It's not the size of the DEM file you need, but the size of the real thing
in meters (it's easier that way in any case).
For instance, if the terrain is supposed to be 2000m (width) x 2000m
(length) x 2341m (maximum height), the scaling should be <2000,2341,2000>.
Of course, this assumes that the maximum height in the DEM file is
translated as pure white in the height field, which may not be the case. It
really depends on how the heights are converted, i.e. what would be the
white value of a map from the Netherlands vs a map of the Himalayas...

G.


-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters

Post a reply to this message

"Gilles Tran" <gitran_nospam_@wanadoo.fr> wrote:

> news:web.41207b10e6dfd373e5987fa60@news.povray.org...
>
> > I think it's OK also because I set :
> > - DEM_size_x and DEM_size_z accordingly to the size of my DEM file,
> > - DEM_size_y is set to 65535 as I use 16bits PNG file.
>
> It's not the size of the DEM file you need, but the size of the real thing
> in meters (it's easier that way in any case).

Sure.

> For instance, if the terrain is supposed to be 2000m (width) x 2000m
> (length) x 2341m (maximum height), the scaling should be <2000,2341,2000>.

My DEM file have a resolution of 75m by pixel and my files is 512x512
pixels.
So I set DEM_size_x and DEM_size_z to 512 * 75 = 38400.

> Of course, this assumes that the maximum height in the DEM file is
> translated as pure white in the height field, which may not be the case. It
> really depends on how the heights are converted, i.e. what would be the
> white value of a map from the Netherlands vs a map of the Himalayas...

The value of each dot of the DEM file is the real altitude in meters. For
example, Annecy's lac altitude is 444m so, the value of this point in 444
(red*256 + green).
As the maximum value for this 16bits value is 65535, I put 65535 as Y scale
....

Laurent

Post a reply to this message

news:web.4120bd99e6dfd373e5987fa60@news.povray.org...
> The value of each dot of the DEM file is the real altitude in meters. For
> example, Annecy's lac altitude is 444m so, the value of this point in 444
> (red*256 + green).
> As the maximum value for this 16bits value is 65535, I put 65535 as Y
scale

Well, then if you're sure the conversion is correct (you can check this by
adding a vertical cylinder of known height to the scene) so it's just a
matter of finding the right camera position, angle and look_at values. It
should be quite accurate, taking into account that the map is low-def.

G.


-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters

Post a reply to this message

Hi All,

First of all, Thanks to every body that help me ;-D

So, I took a photo from a know point :
http://destroyedlolo.homeunix.org:8080/~laurent/dscf0001.jpg

and I make trace :
http://destroyedlolo.homeunix.org:8080/~laurent/angle_30.jpg (with "angle
30")
http://destroyedlolo.homeunix.org:8080/~laurent/sans_angle.jpg (without
"angle")

and ... it's ok : I was confused because, due to numerical artefacts, cliffs
aren't "in the same place" and I though the 1st plan mountain was closer
:-(

Anyway, I'll make some other comparaison from a different point of view when
the weather will be better.

Thanks to all replied to me :-)

Laurent

Post a reply to this message