Hi,

back in 2011 I asked a question about if it was possible to simulate the amount
of light falling on a camera array:
<web.4dded11b6317d989bf90049a0@news.povray.org>

the advice about how to use radiosity to do this was very helpful and I made
some good progress with modelling. However, I have recently spotted something a
bit strange with results. When I have an object emitting onto the focal plane
array (a diffuse reflecting slab) I get the expected results so long as the
object is not too large. However, for larger sizes I get abnormal results.

I have attached a simplified version of code here which shows the problem. The
emitting object is h away from the FPA (h=19) and its diameter is d=15. The
brightness of the pixels in the centre of FPA is exactly correct from radiometry
theory. However, the fall off in intensity towards the edge of array is not
quite steep enough.

If I reduce the size of the emitter to say d=5 I get expected fall off towards
edge of FPA. It seems that the problem occurs for larger emitter sizes.

Any thoughts on what might be causing the effect and if it can be fixed in my
simulation?

Thanks

////sample code//////

#local p_start  = 64/image_width;
#local p_end_final = 8/image_width;

#local final_eb  = 0.05;
#local final_count = 100000;

global_settings{

assumed_gamma 1.0

radiosity
{
    pretrace_start p_start
    pretrace_end   p_end_final

    count final_count
    low_error_factor .05

    nearest_count 1

    error_bound final_eb

    recursion_limit 10

    minimum_reuse 0.00001
    gray_threshold 1
    normal on

    adc_bailout 0.01/50

    }
}

#include "colors.inc"
#include "textures.inc"

#declare h=19; // distance of emitting disc from FPA
#declare d=15; // diameter of emitting disc

#declare ahor=10.24; // width of FPA
#declare aver=ahor/1.25; // height of FPA
//image set up as 640x512, No AA

#declare p=1; // position of camera in front of array
#declare FoVortho =  degrees(2*atan(ahor/(2*p)));       // FoV of orthographic
cam used to view FPA

#declare radianceScene = 50;  // brightness (radiance) of emitting disc

camera {orthographic location <0,0,p> look_at <0,0,0> angle FoVortho}

// the FPA
box {<-ahor/2,-aver/2,-0.5>,<ahor/2,aver/2,0.0>
    texture {
      finish { emission 0 ambient 0 diffuse 0.1 reflection{0} conserve_energy}
pigment {color rgbt<1,1,1,0>}
    }
    scale 1.0 rotate <0,0,0> no_radiosity
}

// emitting disc
disc {<0,0,0>,<0,0,1>, d/2, rotate <0,0,0> translate <0,0,h> pigment {color
rgb<1,1,1>} finish {emission radianceScene ambient 0.0 diffuse 0.0 specular
0.0}}

Post a reply to this message

an update to this problem:

I thought the results were correct for small aperture, however if i run with
more accurate radiosity settings i can see that the same problem is there too.

My intention is that the image created is a greyscale proportional to the
irradiance on each pixel. However, the grey level is ~2% too bright at the edges
of the array compared to theory.

cheers

Post a reply to this message