System and method for measuring image quality

• US 8,494,251 B2
• Filed: 01/12/2009
• Issued: 07/23/2013
• Est. Priority Date: 01/11/2008
• Status: Active Grant

• Current Assignee: SRI International, Inc.
• Original Assignee: SRI International, Inc.
• Inventors: Camus, Theodore Armand
• Primary Examiner(s): Abdi, Amara
• Application Number:
  US12/352,192
• Publication Number:
  US 20090180682A1
• Time in Patent Office: 1,653 Days
• Field of Search: 382/154, 345/419, 348/140, 348/699, 348/E7.085, 712/11, 712/34, 712/35
• US Class Current: 382/154
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 2207/10021 Stereoscopic video; Stereoscopic image sequence
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: H04N 13/239 using two 2D image sensors having a relative position equal to or related to the interocular distance H04N13/243 takes precedence
• CPC Class Codes: H04N 13/296 Synchronisation thereof; Control thereof

Abstract

The present invention provides an improved system and method for measuring quality of both single and stereo video images. The embodiments of the present invention include frequency content measure for a single image or region-of-interest thereof and disparity measure for stereo images or region-of-interest thereof.

Claims

1. A computer-implemented method for measuring Image quality as a determination of eligibility of images to be used for additional processing, comprising:

receiving at least two input images of a region of interest;

determining values of image frequency content of the at least two input images by computing an adjacent pixel difference for each valid pixel in the region of interest, summing the adjacent pixels'"'"' absolute differences of each of the at least two input images and computing the average adjacent pixel difference within the region of interest;

computing a stereo disparity image from the at least two input images if the values are within a specific threshold using the average adjacent pixel difference.

2. The method of claim 1 wherein the at least two input images are original images received from an image sensor and every pixel in the region of interest is considered valid.

3. The method of claim 1 further comprising converting the at least two input images into a binarized disparity image by a binarized disparity module.

4. The method of claim 1 wherein said stereo disparity image is computed from a stereo algorithm, said stereo algorithm explicitly labels each output pixel in the region of interest as valid or invalid.

5. The method of claim 1 wherein each of the at least two input images in the region of interest comprise an entire image.

6. The method of claim 1 wherein each of the at least two input images in the region of interest comprise a left region of interest, a central region of interest and a right region of interest.

7. The method of claim 1 wherein the stereo disparity image is computed if the average adjacent pixel difference of the original image is within a specific threshold.

8. The method of claim 1 comprising converting the stereo disparity image into a binarized disparity image.

9. The method of claim 3 comprising forwarding the disparity image for additional processing if the average adjacent pixel difference of the binarized disparity image is within a specific threshold.

19. the method of claim 1 further comprising:

processing the stereo disparity image to determine eligibility of the stereo disparity image for additional processing comprising at least object detection; and

forwarding the stereo disparity image for the additional processing if the stereo disparity image is determined to be eligible.

10. A computer-implemented system for measuring image quality as a determination of eligibility of images to be used for additional processing, comprising:

an average adjacent pixel difference module to determine values of image frequency content of at least two input images by computing an adjacent pixel difference for each valid pixel in the region of interest, summing the adjacent pixels'"'"' absolute differences of each of the at least two input images and computing the average adjacent pixel difference within the region of interest;

a stereo disparity module to compute a stereo disparity image from the at least two input images using a stereo algorithm comprising using the average adjacent pixel difference if the values are within a specific threshold.

11. The system of claim 10 further comprising at least one sensor to capture the at least two input images in a region of interest.

12. The system of claim 11 wherein the at least two input images are original images received from the sensor and every pixel in the region of interest is considered valid.

13. The system of claim 12 wherein said stereo algorithm explicitly labels each output pixel in the region as valid or invalid.

14. The system of claim 10 further comprising a binarized disparity module to convert the stereo disparity image into a binarized disparity image.

15. The system of claim 10 wherein each of the at least one two input images in the region of interest comprise an entire image.

16. The system of claim 10 wherein each of the at least two input images in the region of interest comprise a left region of interest, a central region of interest and a right region of interest.

17. The system of claim 10 wherein the stereo disparity image is computed if the average adjacent pixel difference of the original image is within a specific threshold.

18. The system of claim 14 wherein the stereo disparity image is forwarded for additional processing if the average adjacent pixel difference of the binarized disparity image is within a specific threshold.

20. the system of claim 10, further comprising:

an object detection module to process the stereo disparity image to determine eligibility of the stereo disparity image for additional processing, comprising at least object detection, and to forward the stereo disparity image for the additional processing if the stereo disparity image is determined to be eligible.