Camera simulation system

• US 6,028,606 A
• Filed: 08/01/1997
• Issued: 02/22/2000
• Est. Priority Date: 08/02/1996
• Status: Expired due to Fees

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: Kolb, Craig E., Hanrahan, Patrick M., Mitchell, Donald P.
• Primary Examiner(s): Powell, Mark R.
• Assistant Examiner(s): HARRISON, CHANTE E
• Application Number:
  US08/904,978
• Time in Patent Office: 935 Days
• Field of Search: 345/421, 345/426, 345/427, 345/433, 345/419, 356/419, 356/3.02, 378/62, 378/4, 382/102, 382/255, 382/154, 382/106, 382/302, 348/36, 351/205, 351/211-214
• US Class Current: 345/419
• CPC Class Codes: G06T 15/06 Ray-tracing
• CPC Class Codes: G06T 15/50 Lighting effects

Abstract

A physical camera is modeled to render an image in a computer graphics system. When given the manufacturer'"'"'s specifications of the physical camera'"'"'s lenses, including the dimensions and indices of refraction of its lenses, stops, and shutter characteristics, the location of the film surface relative to the lens system, and the orientation of the camera within the scene, the invention accurately and efficiently mimics the physical principles of image formation creating an image which approximates an image produced by the physical camera. The procedure comprises four main elements: (1) the geometric relationships between the lens system, object, and film plane are modeled by precise placement and movement of lens elements, (2) image geometry is computed by using principles of geometric optics, (3) an exit pupil is calculated in order to define a region for efficiently sampling rays, (4) the image irradiance, or exposure at a pixel, is computed according to radiometric principles.

Claims

1. A computer implemented method for rendering a three dimensional scene to generate a two dimensional image, the method comprising the steps of:

(a) computing an exit pupil of a lens system;

(b) selecting a ray that passes through the exit pupil to an image point on an image plane;

(c) tracing the ray from the three dimensional scene through the lens system to the image point;

(d) calculating a radiance value for the ray;

wherein the calculation of the radiance value for the ray in step (d) comprises setting the radiance value to zero if the ray is vignetted;

(e) repeating steps (b)-(d) to obtain a plurality of radiance values; and

(f) combining the plurality of radiance values to generate the two dimensional image on the image plane.

2. The method of claim 1 wherein the calculation of the radiance value for the ray in step (d) comprises calculating an attenuation due to losses in the lens system.

3. The method of claim 1 wherein the ray tracing in step (c) comprises the calculation of a 4×

4 projective transformation corresponding to a thick lens approximation of the lens system.

4. The method of claim 1 wherein steps (c) and (d) are performed in dependence upon a wavelength of the ray, whereby rays of differing wavelength may be selected in step (b).

5. The method of claim 1 wherein the combining of the radiance values in step (f) comprises evaluating, for each image point, an approximation to an integral of radiances over the exit pupil.

6. The method of claim 1 wherein the selecting of the ray in step (b) is performed in accordance with a measure preserving mapping of a square to a disk.

7. A computer implemented method for rendering a synthetic two dimensional image of a three dimensional scene in a manner that accurately simulates a physical lens system, the method comprising the steps of:

(a) computing an exit pupil of the lens system;

(b) selecting a ray that passes through the exit pupil from a point on the image surface;

(c) tracing the ray through the lens system to determine an ultimate direction of the ray in object space;

(d) repeating steps (b)-(c) to obtain a first plurality of directions for a first plurality of rays;

(e) computing from the first plurality of rays and the first plurality of directions a selected second plurality of rays and a second plurality of directions;

wherein step (e) includes the use of a thick lens approximation;

(f) calculating radiance values for the second plurality of rays; and

(g) combining the plurality of radiance values to compute an image value at each pixel.

8. The method of claim 7, wherein step (f) comprises setting the radiance value to zero when a ray is vignetted by a shutter, aperture, stop, or other part of the camera system.

9. The method of claim 7, wherein attenuation in the lens system is accounted for in step (f).

10. The method of claim 7, wherein wavelength is accounted for in steps (c)-(g), whereby rays of different wavelengths may be selected in step (b).

11. The method of claim 7, wherein (g) comprises weighting the radiance values associated with rays in close proximity identically.

12. The method of claim 7, wherein (g) includes the use of the differential form factor from a point to a disk.

13. The method of claim 7, wherein (g) includes computing a separate weighting value for each radiance value.

14. The method of claim 7, wherein the two sets of rays and directions in step (e) are identical.

15. The method of claim 7, wherein step (e) includes the use of an approximating function.

16. The method of claim 7, wherein step (e) includes the use of a function that interpolates the plurality of ray directions.

17. The method of claim 7, wherein step (e) includes the use of a linear transformation.

18. The method of claim 7, wherein step (f) further comprises the interpolation of vignetting information.

19. The method of claim 7, wherein rays in are selected using pseudo-random sampling.

20. The method of claim 7, wherein rays are selected using stratified sampling.

21. The method of claim 7, wherein rays are selected by stratifying the pupil and image dimensions independently.

22. The method of claim 7, wherein rays are selected using quasirandom sampling.

23. The method of claim 7, wherein the mapping from a square domain to the domain of the pupil is measure-preserving.

24. The method of claim 7, wherein rays are selected using importance sampling.

25. A computer implemented method for rendering a three dimensional scene to generate a two dimensional image, the method comprising the steps of:

(a) computing an exit pupil of a lens system, wherein the exit pupil defines a boundary between rays blocked by the lens system and rays passing through the lens system;

(b) selecting a ray that passes through the exit pupil to an image point on an image plane;

(c) tracing the ray from the three dimensional scene through the lens system to the image point;

(d) calculating a radiance value for the ray;

(e) repeating steps (b)-(d) to obtain a plurality of radiance values; and

(f) combining the plurality of radiance values to generate the two dimensional image on the image plane.