Quad-edge video signal detector

• US 4,656,507 A
• Filed: 04/10/1984
• Issued: 04/07/1987
• Est. Priority Date: 04/10/1984
• Status: Expired due to Fees

• Current Assignee: Motion Analysis Systems Incorporated
• Original Assignee: Motion Analysis Systems Incorporated
• Inventors: Greaves, John O. B., Warrender, David J.
• Primary Examiner(s): Groody, James J.
• Assistant Examiner(s): Toth, E. Anne
• Application Number:
  US06/598,827
• Time in Patent Office: 1,092 Days
• Field of Search: 358/166, 358/96, 358/183, 358/182, 358/284, 340/728, 340/727, 340/747, 382/22, 382/54, 382/8
• US Class Current: 348/26
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/66 of image moments or centre of gravity

Abstract

Apparatus and method for processing a raster scan video signal representing an image and detecting the outline of one or more objects in the image represented by the video signal. The steps of the method include encoding the raster scan video signal as a sequence of binary pixel values. For each of a plurality of the pixels the following steps are performed. It is detected when the binary pixel values representing the pixel and a horizontally adjacent pixel are distinct. It is also detected when both (a) the binary pixel values representing the pixel and a vertically adjacent pixel are distinct, and (b) the binary pixel values representing a pixel horizontally adjacent to said pixel and a pixel vertically adjacent thereto are distinct. The position of the pixel is reported when either of the detecting steps detects distinct binary pixel values. The second detecting step automatically eliminates redundant pixel reporting by eliminating the reporting of double diagonal pixels.

Claims

1. In a method of processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines, the steps of:

encoding said raster scan video signal as a sequence of binary pixel values,

each said binary pixel value representing one pixel of said image, each said video image scan line being encoded as a multiplicity of binary pixels values representing adjacent pixels, said image being represented by a sequence of a multiplicity of encoded video image scan lines; and

for each of a plurality of said pixels;

detecting when the binary pixel values representing said pixel and a horizontally adjacent pixel are distinct;

detecting when both (a) the binary pixel values representing said pixel and a vertically adjacent pixel are distinct, and (b) the binary pixel values representing a pixel horizontally adjacent to said pixel and a pixel vertically adjacent thereto are distinct; and

reporting said pixel as an outline pixel when either of said detecting steps detects distinct binary pixel values.

2. A method as set forth in claim 1, wherein said encoding and detecting steps are repeated for every pixel of the video signal except those pixels in the first of said sequence of encoded video image scan lines and the first pixel in each said encoded video image scan line.

3. A method as set forth in claim 1,

wherein each said binary pixel value represents either a light or a dark pixel, and adjacent light and dark pixels represent edge transitions between visually distinct portions of said image;

said first and second detecting means include means for detecting four distinct edge transitions in said image, including horizontal light-to-dark and dark-to-light edge transistions and vertical light-to-dark and dark-to-light edge transitions;

said method further including the step of

selecting which of said four edge transitions to detect; and

wherein said reporting step reports said pixel only when at least one of said selected edge transitions is detected.

4. In a system for processing a raster scan video signal representing an image,

encoding means for encoding a raster scan video signal as a sequence of binary pixel values;

first detecting means for detecting when the binary pixel values representing a selected pixel and a horizontally adjacent pixel are distinct;

second detecting means for detecting when both (a) the binary pixel values representing a selected pixel and a vertically adjacent pixel are distinct, and (b) the binary pixel values representing a pixel horizontally adjacent to said selected pixel and a pixel vertically adjacent thereto are distinct; and

means for reporting said pixel as an outline pixel when either of said detecting means detects distinct binary pixel values.

5. A system as set forth in claim 4,

wherein each said binary pixel value represents either a light or a dark pixel, and adjacent light and dark pixels represent edge transitions between visually distinct portions of said image;

said first and second detecting means include means for detecting four distinct edge transitions in said image, including horizontal light-to-dark and dark-to-light edge transitions and vertical light-to-dark and dark-to-light edge transitions;

said system further including

means for selecting which of said four edge transitions to detect; and

wherein said reporting means reports said selected pixel at least only when one of said selected edge transitions is detected.

6. A system as set forth in claim 4, wherein

said second detecting means includes line buffer means for storing said binary pixel values representing at least one raster scan line.

7. A system as set forth in claim 4, wherein

each edge of a visually distinct portion of said image is represented by horizontally, vertically, or diagonally adjacent pixels having distinct pixel values;

said first and second detecting means include read only memory means for detecting each said edge of a visually distinct portion of said image.

8. A system as set forth in claim 7,

wherein each said binary pixel value represents either a light or a dark pixel, and adjacent light and dark pixels represent edge transitions between visually distinct portions of said image;

said read only memory means includes means for detecting four distinct edge transitions in said image, including horizontal light-to-dark and dark-to-light edge transitions and vertical light-to-dark and dark-to-light edge transitions;

said system further including

means coupled to said read only memory means for selecting which of said four edge transitions to detect;

wherein said reporting means reports said selected pixel only when at least one of said selected edge transitions is detected.

9. In a system for processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines,

encoding means for receiving said raster scan video signal and encoding said raster scan video signal as a sequence of quantized pixel values as said video signal is received,

each said quantized pixel value representing one pixel of said image, each said video image scan line being represented by a multiplicity of adjacent pixels, said image being represented by a sequence of a multiplicity of quantized video image scan lines;

buffer means for storing said quantized pixel values for each said video image scan line as the corresponding portion of said video signal is encoded;

first detecting means for detecting, as said video signal is encoded, when the quantized pixel value representing a selected pixel and the quantized pixel value of the previously quantized pixel in the same video image scan line are sufficiently distinct to represent the edge between visually distinct portions of said image;

second detecting means for detecting, as said video signal is encoded, when both (a) the quantized pixel value representing said selected pixel and the quantized pixel value of the adjacent pixel in the previously quantized video image scan line, which was stored in said buffer, are sufficiently distinct to represent the edge between visually distinct portions of said image, and (b) the quantized pixel value representing a pixel adjacent to and in the same video image scan line as said selected pixel and the quantized pixel value of the pixel adjacent thereto in the previously quantized video image scan line, which was stored in said buffer, are sufficiently distinct to represent the edge between visually distinct portions of said image; and

means for reporting said selected pixel as an outline pixel when either of said detecting means detects distinct pixel values.

10. In a method of processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines,

the steps of:

receiving said raster scan video signal and encoding said raster scan video signal as a sequence of quantized pixel values as said video signal is received,

each said quantized pixel value representing one pixel of said image, each said video image scan line being represented by a multiplicity of adjacent pixels, said image being represented by a sequence of a multiplicity of quantized video image scan lines; and

for each of a multiplicity of said pixels in a multiplicity of said video image scan lines, detecting the edge between visually distinct portions of said image, by performing the following steps as said video signal is encoded;

detecting when the quantized pixel value representing said pixel and the quantized pixel value of the previously quantized pixel in the same video image scan line are sufficiently distinct to represent the edge between visually distinct portions of said image;

detecting when both (a) the quantized pixel value representing said pixel and the quantized pixel value of the adjacent pixel in the previously quantized video image scan line, which was stored in said buffer, are sufficiently distinct to represent the edge between visually distinct portions of said image, and (b) the quantized pixel value representing a pixel adjacent to and in the same video image scan line as said pixel and the quantized pixel value of the pixel adjacent thereto in the previously quantized video image scan line, which was stored in said buffer, are sufficiently distinct to represent the edge between visually distinct portions of said image; and

reporting said pixel as an outline pixel when either of said detecting steps detects distinct pixel values.

11. The method of claim 10, including the steps of:

providing a buffer for storing said quantized pixel values; and

storing in said buffer said quantized pixel values for each said video image scan line as the corresponding portion of said video signal is encoded.

12. In a system for processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines,

encoding means for receiving said raster scan video signal and encoding said raster scan video signal as a sequence of binary pixel values as said video signal is received,

each said binary pixel value representing one pixel of said image, each said video image scan line being encoded as a multiplicity of binary pixels values representing adjacent pixels, said image being represented by a sequence of a multiplicity of encoded video image scan lines;

buffer means for storing said binary pixel values for each said video image scan line as the corresponding portion of said video signal is encoded;

first detecting means for detecting, as said video signal is encoded, when the binary pixel value representing a selected pixel and the binary pixel value of the previous pixel in the same video image scan line are distinct;

second detecting means for detecting, as said video signal is encoded, when both (a) the binary pixel value representing said selected pixel and the binary pixel value of the adjacent pixel in the previous video image scan line, which was stored in said buffer, are distinct, and (b) the binary pixel value representing a pixel adjacent to and in the same video image scan line as said selected pixel and the binary pixel value of the pixel adjacent thereto in the previous video image scan line, which was stored in said buffer, are distinct; and

means for reporting said selected pixel as an outline pixel when either of said detecting means detects distinct pixel values.

13. In a method of processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines,

the steps of:

receiving said raster scan video signal and encoding said raster scan video signal as a sequence of binary pixel values as said video signal is received,

each said binary pixel value representing one pixel of said image, each said video image scan line being encoded as a multiplicity of binary pixels values representing adjacent pixels, said image being represented by a sequence of a multiplicity of encoded video image scan lines;

for each of a multiplicity of said pixels in a multiplicity of said video image scan lines, detecting the edge between visually distinct portions of said image, by performing the following steps as said video signal is encoded;

detecting when the binary pixel value representing said pixel and the binary pixel value of the previous pixel in the same video image scan line are distinct;

detecting when both (a) the binary pixel value representing said pixel and the binary pixel value of the adjacent pixel in the previous video image scan line are distinct, and (b) the binary pixel value representing a pixel adjacent to and in the same video image scan line as said pixel and the binary pixel value of the pixel adjacent thereto in the previous video image scan line are distinct; and

reporting said pixel as an outline pixel when either of said detecting steps detects distinct pixel values.

14. The method of claim 13, including the steps of:

providing a buffer for storing said binary pixel values; and

storing said binary pixel values for each said video image scan line substantially immediately after said step of encoding the corresponding portion of said raster scan video signal.

15. In a system for processing a raster scan video signal representing an image, said raster scan video signal representing a multiplicity of video image scan lines,

encoding means for encoding a raster scan video signal as a sequence of binary pixel values,

each said binary pixel value representing one pixel of said image, each said video image scan line being encoded as a multiplicity of binary pixels values representing adjacent pixels, said image being represented by a sequence of a multiplicity of encoded video image scan lines;

buffer means for storing said binary pixel values for at least one video image scan line;

read only memory means having at least four address input signals, including

first table means for generating a report signal;

(1) when the binary pixel value representing a selected pixel and the binary pixel value of the previous pixel in the same video image scan line are distinct; and

(2) when both (a) the binary pixel value representing said selected pixel and the binary pixel value of the adjacent pixel in the previous video image scan line are distinct, and (b) the binary pixel value representing a pixel adjacent to and in the same video image scan line as said pixel and the binary pixel value of the pixel adjacent thereto in the previous video image scan line are distinct; and

second table means for generating a horizontal take signal when the binary pixel value representing said selected pixel and the binary pixel value of the adjacent pixel in the previous video image scan line are distinct;

said address input signals for said read only memory means including the binary pixel value of said selected pixel, the binary pixel value of the previous pixel in the same video image scan line as said selected pixel, the binary pixel value of the pixel adjacent to said selected pixel in the video image scan line previous to the video image scan line containing said selected pixel, and said horizontal take signal generated by said second table means for the previous pixel in the same video image scan line as said selected pixel.

16. In the system of claim 15,

wherein each said binary pixel value represents either a light or a dark pixel, and adjacent light and dark pixels represent edge transitions between visually distinct portions of said image;

said edge transitions including four distinct types of edge transitions;

horizontal light-to-dark, and horizontal dark-to-light, vertical light-to-dark and vertical dark-to-light edge transitions;

said system further including

means for selectively enabling the detection of said four types of edge transitions, including means for generating four separate binary address input signals for said read only memory, each address input signal having a value corresponding to whether the detection of a corresponding one of said edge transitions has been enabled;

wherein said first table means of said read only memory detects only selected edge transitions in said image.