Airport surface fitting.

FYI,

I fixed a couple problems with airport surface fitting.

1. Yeah, kick me, I still had some row/column reversal brain damage.  
The nurbs library matrix routines uses M(row,col) while points are (x, 
y, z) and nurbs surface u,v space was backwards from what I thought it 
should be and nothing is really documented.  I started graphing out 
pictures until I got it all sorted out.

2. nurbs surfaces work in u,v space which is defined across the length 
of the surface, not across x, y.  I was blindly hoping that things would 
be close enough for our purposes, but the difference can be 
substantial.  I introduced a "solver" to find the right u,v 
corresponding to the desired lon,lat point and this also improves the 
resulting surface accuracy quite a bit.

Curt.

-- 
Curtis Olson   Intelligent Vehicles Lab         FlightGear Project
Twin Cities    [email protected]                  [email protected]
Minnesota      http://www.menet.umn.edu/~curt   http://www.flightgear.org