On Saturday, 30 October 2004 05:41, Curtis L. Olson wrote:
> Hi Paul,
>
> There is no polygon layering with the airports right now ... a vertical
> vector placed anywhere over the airport surface will only intersect one
> polygon.
>
> The airport itself is built out of an "irregular" triangle network,
> although there is a distinct pattern used to build the individual
> runways and lay down the different marking textures.
>
> Airport surfaces ... I'm tired, but I'll take a shot at explaining it.
>
> 1. Find the horizontal bounding box of the airport.
>
> 2. Divide this bounding box up into a "regular grid" ... something like
> one point every 500 or 700 meters (I forget the actual number.)  A finer
> grid gives you more variations, a courser grid gives you a smoother
> surface.
>
> 3. For each point in the course grid, sample a much larger number of
> points from the raw terrain data, then average them to come up with the
> elevation for that point.  I seem to recall doing something brilliant
> relating to just sampling points on or near a runway so odd terrain
> between  runways doesn't throw off the whole process.
>
> 4.  Add some contraints to the course grid ... limit the overall amount
> of elevation change across the grid, also limit the maximum slope
> between adjacent grid points.
>
> 5.  Fit a nurbs surface to the adjusted raw grid.  The nurbs surface
> typically provides nice and relatively natural slopes and hills and
> dips.  Doing this over a course grid filters out all the fine grained
> noise.  Some of the other checks and balances help keep things sane if
> there is weird terrain data thrown into the mix.
>
> 6.  Generate the airport in 2d, then drape it over this nurbs surface.
>
> Regards,
>
> Curt.

Thanks Curt, that helps a lot.
Looks like rewriting/replacing genairports could be a whole lot more work than 
I envisioned. :P

Paul

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to