Multiple frame grabber

• US 8,068,182 B2
• Filed: 06/24/2009
• Issued: 11/29/2011
• Est. Priority Date: 10/12/2004
• Status: Expired due to Fees

• Current Assignee: Gula Consulting Limited Liability Company (Intellectual Ventures LLC)
• Original Assignee: Youliza Gehts BV LLC (Intellectual Ventures LLC)
• Inventors: Huettmann, Dirk, Knoedgen, Horst
• Primary Examiner(s): Kostak; Victor
• Application Number:
  US12/456,925
• Publication Number:
  US 20090268035A1
• Time in Patent Office: 888 Days
• Field of Search: 348/187, 348/714, 348/716, 348/718, 348/218.1, 348/35, 348/36, 348/383, 348/231.99, 348/153, 348/159, 348/571, 348/572
• US Class Current: 348/714
• CPC Class Codes: H04N 17/002 for television cameras
• CPC Class Codes: H04N 5/4448 for frame-grabbing

Abstract

A system is presented that applies M×N×K computational units to calculating image parameters on N picture images captured simultaneously by N digital camera devices, where there are N groups of frame grabber units, each containing M frame grabbers in which there are K computational units. The data operated on by a computational unit is separate and independent from the image data operated on by the other computational units. This results in a performance speedup of M×N×K compared to one computational unit making the same computations. A master frame grabber unit controls the illumination of the N digital camera devices, and synchronizes the illumination with the clocks of the N digital camera devices.

Claims

1. A method of parallel processing a picture image from a digital camera device, comprising:

a) dividing a picture image from a digital camera device into M portions,

b) coupling said M portions to M frame grabber units, in which each frame grabber unit contains K computational units,

c) dividing said M portions into K sub-portions and coupling said K sub-portions to said K computational units,

d) calculating parameters of said picture image with said M×

K computational units,

e) coupling results of said calculations to a host computer.

2. The method of claim 1, wherein said sub-portions are by groups of pixels of said picture image.

3. The method of claim 2, wherein said groups of pixels are by row of pixels of said picture image.

4. The method of claim 2, wherein said groups of pixels are by column of pixels of said picture image.

5. The method of claim 2, wherein said groups of pixel are by clusters of pixels forming a checkerboard configuration.

6. The method of claim 1, wherein said calculating parameters of said picture image is performed by said computational units in parallel yielding a performance speedup of M×

K compared to the performance obtained using one computational unit.

7. A method for parallel processing a picture image by pixel, comprising:

a) coupling a picture image to a frame grabber unit containing K computational units,

b) distributing said picture image to said K computational units by pixel,

c) starting with a first computational unit of said K computational units, coupling a first pixel, continuing to the next computational unit until the last computational unit of said K computational units is coupled with a K pixel, then returning to the first computational unit to load a K+1 pixel and continuing in this fashion until all pixels of said picture image is coupled to the K computational units,

d) commencing calculations of image parameters on the pixels of said picture image as additional pixels are coupled to said computational units,

e) coupling results of said calculations to a host computer to be combined to determine a final result of the calculations of said picture image.

8. The method of claim 7, wherein said pixels are coupled to said K computational units by row of pixels of said picture image.

9. The method of claim 7, wherein said pixels are coupled to said K computational units by column of pixels of said picture image.

10. The image of claim 7, wherein said pixels are coupled by a group of pixels having a checkerboard pattern within said picture image.