binding a mesh to a sphere

[peaceman]

hello,

i am currently working on surface meshes for earth (the alps) and had the problem that for large areas meshes out of
srtm-data doesnt fit because they are "flat" - cylindrical projection. i have get a link to your site:
http://orbiter.dansteph.com/forum/index.php?topic=1675.msg22569#msg22569
but since i cant read french i only can see that someone has maybe solved this problem. would be cool if you can help me a bit.

great thanks

peaceman

[picto]

Hello Peaceman

Welcome here ...
The solution I get with Papyref is this one

Z (in meters) = (7,84/100000000) x (x^2 + y^2) in meters)

Where Z is the hight of the vertice you want to move.
And X,Y the position of the vertice from the CENTER of your mesh ...

I tried it on a mesh of Corsica ...
It works fine but it is very long to edit manually all the vertices and
with such big meshes it doesn't solve some visual bugs especially
when you take a look from a little hight ( altitude ... don't know if it is the good word )
in Orbiter.

And if you use 3DsMax, you can use the " Bend deformer "
after the editing of your mesh with the STRM datas ...
But it is not enough " precise "
What I did is to bend  the whole mesh and manually edit all the
vertices from the coast with Papy's solution ...

It is long .... so long ... and boring ....  

I was thinking of another way to do this but never tried it ...
The idea was to built a " curve plane " and keep it like a tool
and editing it with the STRM datas after.
Using a displacement mapping with a black and white drawing on the mesh.
It's very easy to get such drawings with 3DEM and STRM datas ...
Don't know at this time if it is a good way to do the work ...
If you find a better solution, I will be glad to get infos

Good luck to you ...



Post Edited ( 03-18-06 01:18 )

[peaceman]

hi picto,

first thanks for the fast answer. as i see, the manual editing of the vertexes is one way - but a hard way :-)
short before i read your answer, i have figured out this "bend"-function in max, i must experiment with it a bit to see if
its accurate enough for my goal - havent a coastline so it may be not so tragic.
i also tried this displacement mapping on a sphere, really cool thing if you have a bump map of the whole planet (tried it
with titan), but resolution is limited - you can only produce 200 segments in the sphere.
i have tried to delete the most of the sphere and then to get more vertexes by splitting the triangles with a command, this
still works but the position of the vertexes is regular so the whole thing looks irreal if i put the displacement onto.

But i am still experimenting to find a solution ... maybe in future a srtm2msh converter is available - on orbiter forum
urwumpe has started to program this ...

regards

peaceman

[picto]

I've been reading your posts on M6 ... very interesting !
and took a look to your addons ... Nice work !
I hope Urwumpe will be succesful in programming his strm2msh converter.

see if "bend"-function in max to  accurate enough for my goal
- havent a coastline so it may be not so tragic.

You're right !
The problem I have with the mesh I'm building is to be able to land
everywhere on the coastline ... so the bending of the mesh must be perfect in this case .

I remember of a Olympus Mount addon using the property of Orbiter
 to show the whole mesh even if it is under the surface of the planet.

No visual bugs in this case ...

the position of the vertexes is regular

I can see that we encounter the same problems.
Try to use this.
Triangulated Irregular Network
The only problem is to texture this type of mesh ....  
I mean with orthophotographs ...
I think it is better to keep a square base mesh.

See you



Post Edited ( 03-18-06 11:45 )