r.los: patch to support earth curvature

[neteler]

Based on ideas from r.cva and prototype code from Andrea Antonello I have added earth curvature support in r.los. Please test.

Markus

[neteler]

Any opinions on this?

​http://trac.osgeo.org/grass/attachment/ticket/63/rlos_curv.diff

Markus

[hamish]

r.los/main.c:
@@ -109,9 +114,13 @@
     opt6->type = TYPE_DOUBLE;
     opt6->required = NO;
     opt6->answer = "10000";
-    opt6->options = "0-999999";
+    opt6->options = "0-9999999999";
     opt6->description = _("Maximum distance from the viewing point (meters)");

Why increase that so much? At the top of Everest, the maximum theoretical distance to the horizon would be about 3.86*sqrt(8850) = 350km = 350,000m.

​http://mintaka.sdsu.edu/GF/explain/atmos_refr/horizon.html

make it 500000m?

Should all the new /* try to get the radius ... */ code live behind a check for the -c flag? in case the proj stuff has problems.

Hamish

[hopfgartner]

The observer might be on an airplane or other aircraft vehicel

[hopfgartner]

Replying to hopfgartner:

The observer might be on an airplane or other aircraft vehicel

Earth radius might be the ultimative limit, as far as refraction is not considered.

[pkelly]

I've no comment on the maths of the curvature stuff, but I think the code to find the ellpisoid radius should simply use G_get_ellipsoid_parameters() and all the PROJ_INFO parsing and checking for invalid values etc. is taken care of then.

[neteler]

I have simplified the ellipsoid stuff to:

   if (curvature->answer){
      /* try to get the radius the standard GRASS way from the libs */
      G_get_ellipsoid_parameters (&aa, &e2);
      if (aa == 0) {
          /* since there was a problem, take a hardcoded radius :( */
          G_warning(_("Problem to obtain current ellipsoid parameters, using sphere (6370997.0)"));
          aa = 6370997.00;
      }
      G_debug(3, "radius: %f", aa);
    }

Updated earth curvature patch submitted as [c30614] (​http://trac.osgeo.org/grass/changeset/c30614).

Markus