Camera calibration using off-axis illumination and vignetting effects

• US 7,023,472 B1
• Filed: 04/23/1999
• Issued: 04/04/2006
• Est. Priority Date: 04/23/1999
• Status: Expired due to Fees

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Kang, Sing Bing, Weiss, Richard S.
• Primary Examiner(s): Yenke, Brian P
• Application Number:
  US09/298,372
• Time in Patent Office: 2,538 Days
• Field of Search: 348175-176, 348/180, 348187-188, 348/251, 348/223.1, 358/461, 358/504
• US Class Current: 348/187
• CPC Class Codes: G06T 7/80 Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration

Abstract

An imaging device is calibrated using a flat, featureless surface and uniform illumination, relying on the effect of off-axis illumination and vignetting on the reduction of light into the camera at off-axis angles. The effect of the tilt of the camera is also considered. These effects are used to extract intrinsic camera parameters including focal length, principal point, aspect ratio and skew.

Claims

1. A method for calibrating a camera comprising the steps of:

digitizing an image of a blank textureless surface having a uniform illumination;

from the digitized image, determining a pixel intensity drop off caused by a combination of a vignetting effect and an off-axis pixel projection effect due to camera defects;

from the determined pixel intensity drop off, approximating the vignetting effect and the off-axis pixel projection effect using a modeling equation; and

recovering an intrinsic parameter of the camera other than pixel intensity drop off using substantially only the determined pixel intensity drop off.

2. The method as claimed in claim 1 wherein the step of determining is dependent on a camera tilt effect.

3. The method as claimed in claim 1 further comprising the step of computing the parameters of a model by minimizing the difference between the digitized image and the model.

4. The method as claimed in claim 1 wherein the intrinsic parameter is selected from the group consisting of focal length, principal point, skew and aspect ratio.

5. The method as claimed in claim 1 wherein the intrinsic parameter is focal length.

6. The method as claimed in claim 1 wherein the intrinsic parameter is principal point.

7. The method as claimed in claim 1 wherein the intrinsic parameter is skew.

8. The method as claimed in claim 1 wherein the intrinsic parameter is aspect ratio.

9. A computer program product for calibrating a camera, the computer program product comprising a computer usable medium having computer readable code thereon, including program code which:

retrieves a digitized image of a blank textureless surface having a uniform illumination;

from the digitized image, determines a pixel intensity drop off caused by a combination of a vignetting effect and an off-axis pixel projection effect;

from the determined pixel intensity drop off approximating the vignetting effect and the off-axis pixel projection effect using a modeling equation; and

recovers an intrinsic parameter of the camera other than pixel intensity drop off based on substantially only the determined drop off.

10. The computer program product as claimed in claim 9 wherein the program code computes parameters based on a camera tilt effect.

11. The computer program product as claimed in claim 9 wherein the program code computes parameters of a model by minimizing difference between the digitized image and the model.

12. The computer program product as claimed in claim 9 wherein the intrinsic parameter is selected from the group consisting of focal length, principal point, skew and aspect ratio.

13. The computer program product as claimed in claim 9 wherein the intrinsic parameter is focal length.

14. The computer program product as claimed in claim 9 wherein the intrinsic parameter is principal point.

15. The computer program product as claimed in claim 9 wherein the intrinsic parameter is skew.

16. The computer program product as claimed in claim 9 wherein the intrinsic parameter is aspect ratio.

17. A computer system comprising:

a memory system;

an I/O system connected to the memory system;

a storage device connected to the I/O system; and

a calibration routine located in the memory system responsive to a request for calibrating a camera which;

retrieves a digitized image of a blank textureless surface having a uniform illumination;

from the digitized image, determines a pixel intensity drop off caused by a combination of a vignetting effect and an off-axis pixel projection effect;

from the determined pixel intensity drop off, approximating the vignetting effect and the off-axis pixel projection effect using a modeling equation; and

recovers an intrinsic parameter of the camera other than pixel intensity drop off based on substantially only the determined drop off.

18. The computer system as claimed in claim 17 wherein the calibration routine determines parameters dependent on a camera tilt effect.

19. The computer system as claimed in claim 17 wherein the calibration routine computes parameters of a model stored in the storage device, by minimizing difference between the digitized image and the model.

20. The computer system as claimed in claim 17 wherein the intrinsic parameter is selected from the group consisting of focal length, principal point, skew and aspect ratio.

21. The computer system as claimed in claim 17 wherein the intrinsic parameter is focal length.

22. The computer system as claimed in claim 17 wherein the intrinsic parameter is principal point.

23. The computer system as claimed in claim 17 wherein the intrinsic parameter is skew.

24. The computer system as claimed in claim 17 wherein the intrinsic parameter is aspect ratio.

25. An apparatus for calibrating a camera comprising:

means for digitizing an image of a blank textureless surface having a uniform illumination;

means for determining a pixel intensity drop off in the digitized image caused by a combination of a vignetting effect and an off-axis pixel projection effect;

means for approximating the vignetting effect and the off-axis pixel projection effect in the digitized image using a modeling equation; and

means for recovering an intrinsic parameter of the camera other than pixel intensity drop off using substantially only the determined pixel intensity drop off.

26. The apparatus as claimed in claim 25 wherein the means for computing computes parameters based on a camera tilt effect.

27. The apparatus as claimed in claim 25 wherein the means for computing further comprises means for computing parameters of a model by minimizing difference between the digitized image and the model.

28. The apparatus as claimed in claim 25 wherein the intrinsic parameter is selected from the group consisting of focal length, principal point, skew and aspect ratio.

29. The apparatus as claimed in claim 25 wherein the intrinsic parameter is focal length.

30. The apparatus as claimed in claim 25 wherein the intrinsic parameter is principal point.

31. The apparatus as claimed in claim 25 wherein the intrinsic parameter is skew.

32. The apparatus as claimed in claim 25 wherein the intrinsic parameter is aspect ratio.

33. An apparatus for calibrating a camera comprising:

a retrieval routine which retrieves a digitized image of a blank textureless surface having a uniform illumination;

a routine which determines a pixel intensity drop off in the digitized image caused by a combination of a vignetting effect and an off-axis pixel projection effect;

an approximating routine which approximates the vignetting effect and the off-axis pixel projection effect in a digitized image using a modeling equation; and

a parameter computing routine which recovers an intrinsic parameter of the camera other than the pixel intensity drop off using substantially only the determined pixel intensity drop off.

34. The apparatus as claimed in claim 33 wherein the parameter computing routine further comprises a model routine which computes parameters of a model by minimizing difference between the digitized image and the model.

35. A method for calibrating a camera comprising the steps of:

digitizing an image of a blank textureless surface having a uniform illumination;

from the digitized image, determining a pixel intensity drop off caused by a combination of a vignetting effect and an off-axis pixel projection effect;

from the determined pixel intensity drop off, approximating the vignetting effect and the off-axis pixel projection effect using a modeling equation; and

recovering focal length of the camera using substantially only the determined pixel intensity drop off.

36. The method of claim 35 wherein the off-axis pixel projection effect is dependent on the focal length of the camera.

37. The method of claim 36 wherein illuminance of an off-axis image point in the digitized image is proportional to the focal length and distance of the point from the center of the digitized image on the optical axis.

38. A method for calibrating a camera comprising the steps of:

digitizing an image of a blank textureless surface having a uniform illumination;

from the digitized image, determining pixel intensity drop off caused by a combination of a reduction in illumination of image points at the edge of the digitized image and a variation in illumination across the field of view in proportion to the fourth power of the cosine of an angle between a light ray and an optical path;

from the determined pixel intensity drop off approximating the reduction in illumination of image points at the edge of the digitized image and the variation in illumination across the field of view in proportion to the fourth power of the cosine of an angle between a light ray and an optical path; and

recovering an intrinsic parameter of the camera other than pixel intensity drop off using substantially only the determined pixel intensity drop off.

39. A method for providing an estimate for a camera parameter comprising the steps of:

digitizing an image of a blank textureless surface having a uniform illumination;

from the digitized image, determining a pixel intensity drop off caused by a combination of a reduction in illumination of image points at the edge of the digitized image and a variation in illumination across the field of view in proportion to the fourth power of the cosine of an angle between a light ray and an optical path;

from the determined pixel intensity drop off, approximating the reduction in illumination of image points at the edge of the digitized image and the variation in illumination across the field of view in proportion to the fourth power of the cosine of an angle between a light ray and an optical path, and

recovering an intrinsic parameter of the camera other than pixel intensity drop off using substantially only the determined pixel intensity drop off.

40. The method of claim 39, wherein the intrinsic parameter is recovered for use by an image based rendering application.