[Radiance-general] "Sensor Surfaces"

[Thomas Bleicher]

Achim.

I'm afraid I don't quite understand what you want to do.

On Wed, Nov 25, 2009 at 7:40 AM, Achim Geissler
<achim.geissler at intergga.ch> wrote:
> The 'falsecolor' sounds quite interesting, but as far as I can tell it does
> not allow "automatic" evaluation, i.e. create a file / files that can be
> somehow scanned for "hit" or "no hit" per surface. My problem is, that
> visually evaluating images after the simulation does not seem practical due
> to form / size / repetitions.

I recently implemented a script that checks for a hit and miss
of the sky at possible sun positions. I just used the rtrace output
option to produce the material name of the hit surface.

Example input:
0 0 0 0 0 1
0 0 0 0 1 1
0 0 0 1 1 1

Example output:
skyglow
grey_40
grey_40

If you use a unique material name you can afterwards parse the
output with a simple awk one-liner to get the number of hits. Is
that more what you need?

> And I'm afraid I didn't understand if the stencil method could really help,
> but it seems evaluation (or result) is still a "picture" - again not a "hit"
> or "no hit" information for a given surface.

The output is a picture if you want rtrace to produce a picture.
It only depends on the options for rtrace. The nice thing about
the stencil method is that it generates ray origins and directions
over a surface, instead of a fixed size grid.

> Something like
> http://www.radiance-online.org/radiance-workshop7/Content/Augsburger/Germain
> AugsburgerPresentation.pdf, page 8, but with "continuous" and not
> "discrete" sensors that can be evaluated for hit-or-miss somehow is what I
> have in mind.

Stencil could create the x,y,z,dx,dy,dz coordinates but only
for one half of the cylinder (the one facing you). And the
resolution would degrade towards the edges.

Perhaps you could outline your problem. Someone might have
another idea.

Regards,
Thomas