musgrave-forest@CS.YALE.EDU (F. Ken Musgrave) (11/17/90)
Here's a question for all you people out there, one that I'm feeling too lazy to go out and look up the answer to: Am I correct in a vague memory I seem to have, that the (Earth's) moon is supposedly a near-ideal Lambertian reflector? I was recently putting a crescent moon into one of my landscape images, and found that I had to use a gamma of around 16 on the lunar surface, in order to get the terminator on the crescent moon to look right. There's somthing fishy here, methinks. Ken -- Ken Musgrave musgrave-forest@yale.edu Yale U Depts of CS and Math (203) 432-4016 Box 2155 Yale Station "But Mr. Natural! is there any future?!?" New Haven, CT 06520 "Not yet."
shirley@iuvax.cs.indiana.edu (peter shirley) (11/18/90)
musgrave-forest@CS.YALE.EDU (F. Ken Musgrave) writes: > Am I correct in a vague memory I seem to have, that the (Earth's) moon is >supposedly a near-ideal Lambertian reflector? > There's somthing fishy here, methinks. The moon is not lambertian. If it were, then a full moon would be bright at the center of the disk, and fade into black at the edges. Instead, a full moon is a fairly uniformly colored disk. If you assume the lambertian surface has a scattering probability kcos(theta), you might guess that the moon is just k (scatters in all directions above the surface with equal probabilities). I think this will work for a simple model. pete PS-- I saw some angular reflectance curves in some book I can't find. I think it was a heat transfer text. The actual function was pretty funky. PPS-- Nice pictures in this months CG&A! (by FKM)
fournier@cs.ubc.ca (Alain Fournier) (11/19/90)
In article <27331@cs.yale.edu> musgrave-forest@CS.YALE.EDU (F. Ken Musgrave) writes: >... > Am I correct in a vague memory I seem to have, that the (Earth's) moon is >supposedly a near-ideal Lambertian reflector? > >... As this is one of my favourite trick question about reflection, I'll bite. Consider the moon when full (that is the sun -the light source- is in the same direction as the observer -the eye. It is pretty obvious that it appears essentially as a disk, that is the reflected light is about of same luminance all over the visible part of the moon (I ignore the effect of the surface details such as the maria). That shows that it is neither a specular reflector (the centre would be a lot brighter than the periphery, nor a diffuse reflector (the centre would be somewhat brighter than the periphery (try this on your favourite renderer, a perfectly diffuse white sphere illuminated from the direction of the eye). So what's going on. Well, as most of the computer graphics types (us) ignore, the world is not all between totally diffuse and totally specular, there are surfaces outside of this. In the case of the moon, it so happens that a Phong model (using the expression loosely) with an exponent of 0.5 for the cosine of the angle normal/light gives the right appearance of a disk at full moon. I can find exact references back in my office, if anybody is interested. Credit where credit is due: Bob Woodham, of UBC, first pointed that out to me, and has worked on the subject of models for the surface reflectance of the moon (and other objects in the solar system).
mcdonald@aries.scs.uiuc.edu (Doug McDonald) (11/19/90)
In article <10506@ubc-cs.UUCP> fournier@cs.ubc.ca (Alain Fournier) writes: >As this is one of my favourite trick question about reflection, I'll bite. >Consider the moon when full (that is the sun -the light source- is in the same >direction as the observer -the eye. It is pretty obvious that it appears >essentially as a disk, that is the reflected light is about of same >luminance all over the visible part of the moon (I ignore the effect >of the surface details such as the maria). That shows that it is neither >a specular reflector (the centre would be a lot brighter than the periphery, >nor a diffuse reflector (the centre would be somewhat brighter than the >periphery (try this on your favourite renderer, a perfectly diffuse white >sphere illuminated from the direction of the eye). So what's going on. >Well, as most of the computer graphics types (us) ignore, the world is not all >between totally diffuse and totally specular, there are surfaces outside >of this. In the case of the moon, it so happens that a Phong model >(using the expression loosely) with an exponent of 0.5 for the cosine >of the angle normal/light gives the right appearance of a disk at full moon. >I can find exact references back in my office, if anybody is interested. >Credit where credit is due: Bob Woodham, of UBC, first pointed that out to me, >and has worked on the subject of models for the surface reflectance of the >moon (and other objects in the solar system). Remember that the moon has hills and valleys. The hills near the terminator are what one sees, and the actual surface is seen at an angle much less than 90 degrees. The actual surface itself is sort of Lambertian. At least when I actually examined some small (~5 cm) size pieces of the Moon in a lab they looked rather Lambertian. On the spot reports agree. The word "fractal" comes to mind. Doug McDonald
ph@miro.Berkeley.EDU (Paul Heckbert) (11/27/90)
musgrave-forest@CS.YALE.EDU (F. Ken Musgrave) writes: > Am I correct in a vague memory I seem to have, that the (Earth's) moon is >supposedly a near-ideal Lambertian reflector? As others have mentioned, the moon and other dusty surfaces are not Lambertian. Blinn, in the paper cited below, says they follow the Hapke-Irvine illumination model more closely. %A James F. Blinn %T Light Reflection Functions for Simulation of Clouds and Dusty Surfaces %J Computer Graphics (SIGGRAPH '82 Proceedings) %V 16 %N 3 %D July 1982 %P 21-29 %K shading %Z Hapke-Irvine illumination model
leech@homer.cs.unc.edu (Jonathan Leech) (11/28/90)
In article <9237@pasteur.Berkeley.EDU>, ph@miro.Berkeley.EDU (Paul Heckbert) writes: |>As others have mentioned, the moon and other dusty surfaces are not Lambertian. |>Blinn, in the paper cited below, says they follow the Hapke-Irvine |>illumination model more closely. This may also be of interest, although rather more technical than most computer graphics papers. A good deal of work would need to be done to convert the results into a lighting model. Lumme & Bowell, _Radiative Transfer in the Surfaces of Atmosphereless Bodies_, The Astronomical Journal 86:11, November 1981. This covers radiative transfer with multiple scattering in porous, rough surfaces.