Method of filtering an image

• US 8,768,007 B2
• Filed: 03/26/2012
• Issued: 07/01/2014
• Est. Priority Date: 03/26/2012
• Status: Expired due to Fees

• Current Assignee: Joyson Safety Systems Acquisition, LLC
• Original Assignee: TK Holdings Incorporated (Ningbo Junsheng Investment Group Co., Ltd.)
• Inventors: Schamp, Gregory Gerhard
• Primary Examiner(s): Koziol, Stephen R
• Assistant Examiner(s): CHOI, TIMOTHY WING HO
• Application Number:
  US13/429,803
• Publication Number:
  US 20130251193A1
• Time in Patent Office: 827 Days
• Field of Search: 382/103, 382/209, 382/210, 382260-265, 382/274, 382/275
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 2207/20068 Projection on vertical or horizontal image axis
• CPC Class Codes: G06T 2207/20168 Radial search
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/593 from stereo images
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 20/64 Three-dimensional objects

Abstract

For each of a plurality of homogeneity regions in relative pixel space, a deviation associated therewith with respect to a pixel to be filtered is given by a sum of associated difference values, each difference value given by the absolute value of a difference between a value of a pixel to be filtered and that of a neighboring pixel selected in accordance with the selected homogeneity region. The filtered pixel value is responsive to values of the neighboring pixels for the homogeneity region with minimum deviation. The relative pixel locations of each homogeneity region are symmetric relative to a radially-extending axis extending outwards in a corresponding polar direction therealong from a relatively central pixel location to a boundary of the homogeneity region, including all relative pixels intersected by the radially-extending axis, different homogeneity regions being associated with different polar directions.

Claims

1. A method of filtering an image, comprising:

a. identifying a plurality of homogeneity regions in relative pixel space, wherein each homogeneity region of said plurality of homogeneity regions comprises a plurality of active relative pixels located at a corresponding plurality of locations that are located relative to a relatively central pixel within a boundary circumscribing a composite of all of said plurality of homogeneity regions, said plurality of locations of said plurality of active relative pixels are symmetric relative to a corresponding radially-extending axis extending outwards in a corresponding polar direction from the location of said relatively central pixel to said boundary, said plurality of locations of said plurality of active relative pixels include all active relative pixels thereof that are intersected by said radially-extending axis, and each different homogeneity region of said plurality of homogeneity regions corresponds to a different said radially-extending axis extending in a different said polar direction relative to an orientation of said boundary;

b. receiving or determining a location of a pixel of the image to be filtered;

c. for each of said plurality of homogeneity regions;

i. selecting one of said plurality of homogeneity regions;

ii. calculating a deviation associated with said one of said plurality of homogeneity regions, wherein said relatively central pixel of said plurality of homogeneity regions is co-located with said pixel of said image, said deviation is given by a sum of a plurality of associated difference values, and each associated difference value of said plurality of associated difference values is given by the absolute value of a difference between a value of said pixel to be filtered and a value of a neighboring pixel of said image at a relative location corresponding to a corresponding relative location of a selected active relative pixel of said homogeneity region, for each of said plurality of active relative pixels of said one of said plurality of homogeneity regions corresponding to a corresponding plurality of neighboring pixels of said image, wherein each said neighboring pixel of said image is one of said plurality of neighboring pixels of said image;

d. finding at least one said homogeneity region of said plurality of homogeneity regions for which a corresponding said deviation is a minimum;

e. generating a filtered pixel value of said pixel of said image to be filtered, wherein said filtered pixel value is responsive to values of each of said plurality of neighboring pixels of said image to be filtered, and each said neighboring pixel of said plurality of neighboring pixels corresponds to a different said selected active relative pixel of said at least one said homogeneity region for which said corresponding said deviation is said minimum.

2. A method of filtering an image as recited in claim 1, wherein each said polar direction is selected from the group consisting of angles of 45 degrees, 90 degrees, 135 degrees, 215 degrees, 270 degrees and 315 degrees as measured from an X-axis of said relative pixel space, wherein a row dimension of said relative pixel space is aligned with said X-axis.

3. A method of filtering an image as recited in claim 1, wherein said plurality of active relative pixels comprise three relatively-proximal pixels and one relatively distal pixel, each of said three relatively-proximal pixels is either laterally adjacent or diagonally adjacent to said relatively central pixel, and said one relatively distal pixel is either laterally adjacent or diagonally adjacent to, and relatively distal with respect to, one of said three relatively-proximal pixels along said radially-extending axis.

4. A method of filtering an image as recited in claim 1, wherein said filtered pixel value is generated from an average of said values of said plurality of neighboring pixels.

5. A method of filtering an image as recited in claim 4, wherein a count of said plurality of neighboring pixels is equal to a power of two, and said average of said values of said plurality of neighboring pixels is determined from a binary right-shift of a sum of said values of said plurality of neighboring pixels by an amount corresponding to said power of two.

6. A method of filtering an image as recited in claim 1, wherein said filtered pixel value is generated as a median value of said values of said plurality of neighboring pixels.