Method, System and Computer Program Product for Identifying Locations of Detected Objects

• US 20130121562A1
• Filed: 11/13/2012
• Published: 05/16/2013
• Est. Priority Date: 11/11/2011
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Barnum, Peter
• Granted Patent: US 8,831,337 B2
• Time in Patent Office: Days
• US Class Current: 382/154
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06V 20/64 Three-dimensional objects

Abstract

First and second objects are detected within an image. The first object includes first pixel columns, and the second object includes second pixel columns. A rightmost one of the first pixel columns is adjacent to a leftmost one of the second pixel columns. A first equation is fitted to respective depths of the first pixel columns, and a first depth is computed of the rightmost one of the first pixel columns in response to the first equation. A second equation is fitted to respective depths of the second pixel columns, and a second depth is computed of the leftmost one of the second pixel columns in response to the second equation. The first and second objects are merged in response to the first and second depths being sufficiently similar to one another, and in response to the first and second equations being sufficiently similar to one another.

Claims

1. A method performed by at least one device for enhancing accuracy in identifying locations of detected objects within an image, the method comprising:

detecting a first object within the image, wherein the first object includes pixel columns;

fitting an equation to respective depths of the pixel columns, and determining whether a fit is sufficient between the equation and the respective depths of the pixel columns; and

in response to determining that the fit is insufficient between the equation and more than a threshold number of the respective depths of consecutive pixel columns, splitting the first object into at least second and third objects, wherein the second object includes the consecutive pixel columns.

2. The method of claim 1, wherein the image is a stereoscopic image, and comprising:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

in response to the respective disparity values, estimating the respective depths of the pixel columns.

3. The method of claim 1, wherein the respective depths of the pixel columns are estimated depths, and wherein determining whether the fit is sufficient includes:

in response to the equation, computing depths of the pixel columns; and

comparing the estimated depths against the computed depths.

4. The method of claim 3, wherein the comparing includes:

determining whether a threshold is exceeded by differences between the estimated depths and the computed depths.

5. The method of claim 1, wherein the respective depths of the pixel columns are estimated depths of nearest detected objects in the pixel columns, and wherein the first object is a merger of the nearest detected objects.

6. A method performed by at least one device for enhancing accuracy in identifying locations of detected objects within an image, the method comprising:

detecting first and second objects within the image, wherein the first object includes first pixel columns, wherein the second object includes second pixel columns, and wherein a rightmost one of the first pixel columns is adjacent to a leftmost one of the second pixel columns;

fitting a first equation to respective depths of the first pixel columns, and computing a first depth of the rightmost one of the first pixel columns in response to the first equation;

fitting a second equation to respective depths of the second pixel columns, and computing a second depth of the leftmost one of the second pixel columns in response to the second equation; and

merging the first and second objects in response to the first and second depths being sufficiently similar to one another, and in response to the first and second equations being sufficiently similar to one another.

7. The method of claim 6, wherein the image is a stereoscopic image, and comprising:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

in response to the respective disparity values, estimating the respective depths of the first pixel columns and of the second pixel columns.

8. The method of claim 6, wherein the merging includes:

determining whether the first and second depths are sufficiently similar to one another by determining whether a threshold is exceeded by a difference between the first and second depths.

9. The method of claim 8, wherein the threshold is a first threshold, and wherein the merging includes:

determining whether the first and second equations are sufficiently similar to one another by determining whether a second threshold is exceeded by a difference between a constant factor of the first equation and a constant factor of the second equation.

10. The method of claim 9, wherein the first and second equations are linear equations.

11. A system for enhancing accuracy in identifying locations of detected objects within an image, the system comprising:

at least one device for;

detecting a first object within the image, wherein the first object includes pixel columns;

fitting an equation to respective depths of the pixel columns, and determining whether a fit is sufficient between the equation and the respective depths of the pixel columns; and

, in response to determining that the fit is insufficient between the equation and more than a threshold number of the respective depths of consecutive pixel columns, splitting the first object into at least second and third objects, wherein the second object includes the consecutive pixel columns.

12. The system of claim 11, wherein the image is a stereoscopic image, and wherein the at least one device is for:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

, in response to the respective disparity values, estimating the respective depths of the pixel columns.

13. The system of claim 11, wherein the respective depths of the pixel columns are estimated depths, and wherein determining whether the fit is sufficient includes:

in response to the equation, computing depths of the pixel columns; and

comparing the estimated depths against the computed depths.

14. The system of claim 13, wherein the comparing includes:

determining whether a threshold is exceeded by differences between the estimated depths and the computed depths.

15. The system of claim 11, wherein the respective depths of the pixel columns are estimated depths of nearest detected objects in the pixel columns, and wherein the first object is a merger of the nearest detected objects.

16. A system for enhancing accuracy in identifying locations of detected objects within an image, the system comprising:

at least one device for;

detecting first and second objects within the image, wherein the first object includes first pixel columns, wherein the second object includes second pixel columns, and wherein a rightmost one of the first pixel columns is adjacent to a leftmost one of the second pixel columns;

fitting a first equation to respective depths of the first pixel columns, and computing a first depth of the rightmost one of the first pixel columns in response to the first equation;

fitting a second equation to respective depths of the second pixel columns, and computing a second depth of the leftmost one of the second pixel columns in response to the second equation; and

merging the first and second objects in response to the first and second depths being sufficiently similar to one another, and in response to the first and second equations being sufficiently similar to one another.

17. The system of claim 16, wherein the image is a stereoscopic image, and wherein the at least one device is for:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

, in response to the respective disparity values, estimating the respective depths of the first pixel columns and of the second pixel columns.

18. The system of claim 16, wherein the merging includes:

determining whether the first and second depths are sufficiently similar to one another by determining whether a threshold is exceeded by a difference between the first and second depths.

19. The system of claim 18, wherein the threshold is a first threshold, and wherein the merging includes:

determining whether the first and second equations are sufficiently similar to one another by determining whether a second threshold is exceeded by a difference between a constant factor of the first equation and a constant factor of the second equation.

20. The system of claim 19, wherein the first and second equations are linear equations.

21. A computer program product for enhancing accuracy in identifying locations of detected objects within an image, the computer program product comprising:

a tangible computer-readable storage medium; and

a computer-readable program stored on the tangible computer-readable storage medium, wherein the computer-readable program is processable by an information handling system for causing the information handling system to perform operations including;

detecting a first object within the image, wherein the first object includes pixel columns;

fitting an equation to respective depths of the pixel columns, and determining whether a fit is sufficient between the equation and the respective depths of the pixel columns; and

, in response to determining that the fit is insufficient between the equation and more than a threshold number of the respective depths of consecutive pixel columns, splitting the first object into at least second and third objects, wherein the second object includes the consecutive pixel columns.

22. The computer program product of claim 21, wherein the image is a stereoscopic image, and wherein the operations include:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

, in response to the respective disparity values, estimating the respective depths of the pixel columns.

23. The computer program product of claim 21, wherein the respective depths of the pixel columns are estimated depths, and wherein determining whether the fit is sufficient includes:

in response to the equation, computing depths of the pixel columns; and

comparing the estimated depths against the computed depths.

24. The computer program product of claim 23, wherein the comparing includes:

determining whether a threshold is exceeded by differences between the estimated depths and the computed depths.

25. The computer program product of claim 21, wherein the respective depths of the pixel columns are estimated depths of nearest detected objects in the pixel columns, and wherein the first object is a merger of the nearest detected objects.

26. A computer program product for enhancing accuracy in identifying locations of detected objects within an image, the computer program product comprising:

a tangible computer-readable storage medium; and

a computer-readable program stored on the tangible computer-readable storage medium, wherein the computer-readable program is processable by an information handling system for causing the information handling system to perform operations including;

detecting first and second objects within the image, wherein the first object includes first pixel columns, wherein the second object includes second pixel columns, and wherein a rightmost one of the first pixel columns is adjacent to a leftmost one of the second pixel columns;

fitting a first equation to respective depths of the first pixel columns, and computing a first depth of the rightmost one of the first pixel columns in response to the first equation;

fitting a second equation to respective depths of the second pixel columns, and computing a second depth of the leftmost one of the second pixel columns in response to the second equation; and

merging the first and second objects in response to the first and second depths being sufficiently similar to one another, and in response to the first and second equations being sufficiently similar to one another.

27. The computer program product of claim 26, wherein the image is a stereoscopic image, and wherein the operations include:

in response to the stereoscopic image, generating a depth map that assigns respective disparity values to pixel coordinates within the stereoscopic image, wherein the respective disparity values are indicative of respective depths of the pixel coordinates within the stereoscopic image; and

, in response to the respective disparity values, estimating the respective depths of the first pixel columns and of the second pixel columns.

28. The computer program product of claim 26, wherein the merging includes:

determining whether the first and second depths are sufficiently similar to one another by determining whether a threshold is exceeded by a difference between the first and second depths.

29. The computer program product of claim 28, wherein the threshold is a first threshold, and wherein the merging includes:

determining whether the first and second equations are sufficiently similar to one another by determining whether a second threshold is exceeded by a difference between a constant factor of the first equation and a constant factor of the second equation.

30. The computer program product of claim 29, wherein the first and second equations are linear equations.