Simple method for calculating camera defocus from an image scene

• US 7,990,462 B2
• Filed: 10/15/2009
• Issued: 08/02/2011
• Est. Priority Date: 03/16/2006
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.), Sony Electronics Inc. (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.), Sony Electronics Inc. (Sony Group Corp.)
• Inventors: Nakamura, Makibi, Wong, Earl Quong, Kushida, Hidenori
• Primary Examiner(s): Ometz; David L
• Assistant Examiner(s): Berhan; Ahmed A
• Application Number:
  US12/580,217
• Publication Number:
  US 20100033618A1
• Time in Patent Office: 656 Days
• Field of Search: 348/344, 348/345, 348/346, 348/350, 348/353, 348/97, 348/46, 250/201.7
• US Class Current: 348/353
• CPC Class Codes: H04N 23/673 based on contrast or high frequency components of image signals, e.g. hill climbing method

Abstract

An imaging acquisition system that generates a picture depth from an auto focus curve generated from picture of a three dimensional spatial scene is described. The auto focus curve comprises a step edge. The system generates the depth based on the step edge and a reference auto focus normalization curve.

Claims

1. A computerized method comprising:

receiving an auto focus curve that is associated with a picture of a three dimensional spatial scene, wherein the auto focus curves comprises a step edge; and

computing, by a processor in a camera, a picture depth value using the step edge and an auto focus normalization reference curve, wherein the auto focus normalization reference curve includes a plurality of points and correlates the step edge to a known amount of blurring, and wherein the computing a picture depth comprises,

computing a defocus measurement value for a step edge in the auto focus curve using a one dimensional step edge gradient and a step edge center,

deriving a picture blur value using the defocus measurement value and the auto focus normalization reference curve, and

computing the picture depth value using the picture blur value.

2. The computerized method of claim 1, further comprising:

generating a plurality of auto focus normalization reference curves for different step edge angles.

3. The computerized method of claim 2, further comprising:

selecting the auto focus normalization reference curve from a plurality of auto focus normalization reference curves, where the auto focus normalization reference curve is associated with an attribute selected from the group consisting of a step edge angle and a step edge contrast.

4. The computerized method of claim 1, further comprising:

generating a picture depth map by computing a plurality of picture depth values for a plurality of step edges.

5. The computerized method of claim 4, further comprising:

extracting regions of similar depth from the picture depth map using a clustering algorithm.

6. The computerized method of claim 1, wherein the computing the picture depth value further comprises:

detecting the step edge in the auto focus curve;

shifting the step edge to have a baseline of zero;

computing an one dimensional gradient of the step edge; and

locating a step edge center.

7. The computerized method of claim 6, wherein the computing the picture depth value further comprises:

calculating, from the picture blur value, the picture depth value using the equation

8. A non-transitory machine readable storage medium having executable instructions to cause a processor to perform a method comprising:

receiving an auto focus curve, that is associated with a picture of a three dimensional spatial scene, wherein the auto focus curves comprises a step edge; and

computing a picture depth value using the step edge and an auto focus normalization reference curve, wherein the auto focus normalization reference curve includes a plurality of points and correlates the step edge to a known amount of blurring, wherein the computing a picture depth comprises,

computing a defocus measurement value for a step edge in the auto focus curve using a one dimensional step edge gradient and a step edge center,

deriving a picture blur value using the defocus measurement value and the auto focus normalization reference curve, and

computing the picture depth value using the picture blur value.

9. The non-transitory machine readable storage medium of claim 8, further comprising:

generating a plurality of auto focus normalization reference curves for different step edge angles.

10. The non-transitory machine readable storage medium of claim 9, further comprising:

selecting the auto focus normalization reference curve from a plurality of auto focus normalization reference curves, where the auto focus normalization reference curve is associated with an attribute selected from the group consisting of a step edge angle and a step edge contrast.

11. The non-transitory machine readable storage medium of claim 8, further comprising:

generating a picture depth map by computing a plurality of picture depth values for a plurality of step edges.

12. The non-transitory machine readable storage medium of claim 11, further comprising:

extracting regions of similar depth from the picture depth map using a clustering algorithm.

13. The non-transitory machine readable storage medium of claim 8, wherein the computing the picture depth value further comprises:

detecting the step edge in the auto focus curve;

shifting the step edge to have a baseline of zero;

computing an one dimensional gradient of the step edge; and

locating a step edge center.

14. The non-transitory machine readable storage medium of claim 13, wherein the computing the picture depth value further comprises:

calculating, from the picture blur value, the picture depth value using the equation

15. An apparatus comprising:

means for receiving an auto focus curve, that is associated with a picture of a three dimensional spatial scene, wherein the auto focus curves comprises a step edge; and

means for computing a picture depth value using the step edge and an auto focus normalization reference curve, wherein the auto focus normalization reference curve includes a plurality of points and correlates the step edge to a known amount of blurring and the means for computing the picture depth value comprises

means for computing a defocus measurement value for the step edge using the one dimensional step edge gradient and step edge center,

means for deriving a picture blur value using the defocus measurement value and the auto focus normalization reference curve, and

means for computing the picture depth value using the picture blur value.

16. The apparatus of claim 15, further comprising:

means for generating a plurality of auto focus normalization reference curves for different step edge angles.

17. The apparatus of claim 16, further comprising:

means for selecting the auto focus normalization reference curve from a plurality of auto focus normalization reference curves, where the auto focus normalization reference curve is associated with an attribute selected from the group consisting of a step edge angle and a step edge contrast.

18. The apparatus of claim 15, further comprising:

means for generating a picture depth map by computing a plurality of picture depth values for a plurality of step edges.

19. The apparatus of claim 15, wherein the means for computing the picture depth value further comprises:

means for detecting the step edge in the auto focus curve;

means for shifting the step edge to have a baseline of zero;

means for computing an one dimensional gradient of the step edge; and

means for locating a step edge center.

20. A system comprising:

a processor;

a memory coupled to the processor though a bus; and

a process executed from the memory by the processor to cause the processor to receive an auto focus curve, that is associated with a picture of a three dimensional spatial scene, wherein the auto focus curves comprises a step edge; and

compute a picture depth value using the step edge and an auto focus normalization reference curve, wherein the auto focus normalization reference curve includes a plurality of points and correlates the step edge to a known amount of blurring and the process further causes the processor to compute a defocus measurement value for the step edge using the one dimensional step edge gradient and step edge center, derive a picture blur value using the defocus measurement value and the auto focus normalization reference curve, and compute the picture depth value using the picture blur value.

21. The system of claim 20, wherein the process further causes the processor to generate a plurality of auto focus normalization reference curves for different step edge angles.

22. The system of claim 21, wherein the process further causes the processor to select the auto focus normalization reference curve from a plurality of auto focus normalization reference curves, where the auto focus normalization reference curve is associated with an attribute selected from the group consisting of a step edge angle and a step edge contrast.

23. The system of claim 20, wherein the process further causes the processor to generate a picture depth map by computing a plurality of picture depth values for a plurality of step edges.

24. The system of claim 20, wherein the process further causes the processor to detect the step edge in the auto focus curve, shift the step edge to have a baseline of zero, compute an one dimensional gradient of the step edge, and locate a step edge center.