home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!usc!zaphod.mps.ohio-state.edu!rpi!utcsri!dgp.toronto.edu!dret
- Newsgroups: sci.math
- From: dret@dgp.toronto.edu (George Drettakis)
- Subject: Properties of Illumination Functions?
- Message-ID: <1992Dec16.225253.25173@jarvis.csri.toronto.edu>
- Organization: CSRI, University of Toronto
- Date: 17 Dec 92 03:52:53 GMT
- Lines: 86
-
- I sent this to sci.math.research, but I have a suspicion the big
- black net hole ate it. I apologize if you see this twice.
- -----------------------
-
- I have the following problem in computer graphics. Assume a rectangular
- area in the plane z=d, with -xa1 < x1 < xa1 and -ya1 < y1 < ya1.
- This area is assumed to be a diffuse rectangular light source,
- with the surface normal 0, 0, -1.
- Assume also the receiving plane z=0, and the function I(x,y) defined on
- this plane. The function I(x,y) expresses the illumination at this point:
-
- ya1 xa1
- / /
- | | 1
- I(x, y) := | | ----------------------------- dx1 dy1
- | | 2 2 2 2
- / / ((x - x1) + (y - y1) + d )
- - ya1 - xa1
-
- The analytic solution to this integral is as follows:
-
-
- (y - ya1) w1 (y + ya1) w3 (x - xa1) w4
- I(x,y)= - ------------------- + ------------------- - -------------------
- 2 2 1/2 2 2 1/2 2 2 1/2
- (d + (y - ya1) ) (d + (y + ya1) ) (d + (x - xa1) )
-
- (x + xa1) w2
- + -------------------
- 2 2 1/2
- (d + (x + xa1) )
-
-
- where, wi is the angle (in radians) formed by the two consecutive vertices
- of the source and the receiving point.
- (In the figure w1 is the angle < v1 P v2 )
-
- source: v1
- ---------- +
- / / +
- / + / +
- / / + <-angle w1
- ----------+ +
- -ya1 v2 + +
- +ya1 + +
- __+___+________________________________
- / + + P=(x,y,0) receiver plane /
- / /
- / /
- /______________________________________/
-
-
- Each w i is usually expressed as an arccos or an arctan.
- As an example of the difficulty of these expressions, for xa1=ya1=1,
- w1 is:
-
- 2 2
- d + y - 2 y + x
- w1=arccos(-----------------------------------------------------------------)
- 2 2 1/2 2 2 1/2
- (d + y - 2 y + 2 + x - 2 x) (d + y - 2 y + 2 + x + 2 x)
-
-
-
- What I want to show is that the function I(x,y) has only one maximum, and is
- (radially around 0, 0, 0 ) decreasing everywhere else.
- Ideally I want to show this for more general sources, but just the rectangle
- is a good start.
-
- I have some ideas, but before I spend more time on this, I would like to
- know whether this class of functions is well known, and whether there are
- easy ways to prove its properties, or whether I am missing the obvious.
- Derivative analysis seems to be inconclusive, due to the complexity of
- the expressions. (The partials result in very high-degree equations).
-
- Pointers to books, papers etc. are particularily welcome.
- I apologize beforehand if this is a well known problem with well known
- solutions. I dont read this group often.
-
- Thanks,
-
- George Drettakis Dynamic Graphics Project
- Ph.D. Student Dept. of Computer Science
- e-mail dret@dgp.toronto.edu Sandford Fleming Bldg.
- Phone +1 (416) 978-5473 10, Kings College Rd.
- FAX +1 (416) 978-5184 Toronto, Ontario CANADA M5S 1A1
-