Mechanism for tracking colored objects in a video sequence

• US 6,760,465 B2
• Filed: 03/30/2001
• Issued: 07/06/2004
• Est. Priority Date: 03/30/2001
• Status: Expired due to Term

• Current Assignee: Beijing Xiaomi Mobile Software Co., Ltd. (Xiaomi Corp.)
• Original Assignee: Intel Corporation
• Inventors: McVeigh, Jeffrey S, Frazer, Matthew E
• Primary Examiner(s): Johns, Andrew W.
• Assistant Examiner(s): NAKHJAVAN, SHERVIN K
• Application Number:
  US09/822,086
• Publication Number:
  US 20020141615A1
• Time in Patent Office: 1,194 Days
• Field of Search: 382/103, 382/169, 382/172, 348/169, 348/172, 348/208.14, 725/12
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes

Abstract

An automated object tracking system for tracking a colored object through a series of frames of data. The system has a first image source device to provide a data array of pixels of a digital image. A second image source device provides a binary image of the data array. The system utilizes an analysis system to create a first histogram for columns of the binary image and a second histogram for rows of the binary image. Each histogram is thresholded, and after being thresholding, is utilized to estimate the center of a tracked object in the binary image.

Claims

1. An automated object tracking system for tracking a colored object through a series of frames of data, comprising:

a first image source device to provide a data array of pixels of a digital image;

a second image source device to provide a binary image of the data array; and

an analysis system to create a first histogram for columns of the binary image, and a second histogram for rows of the binary image, each of the first histogram and the second histogram being thresholded and filtered to create a first thresholded filtered histogram and a second thresholded filtered histogram, wherein the filtering reduces the effect of noise and reduces the effect of pixels in the data array away from a main concentration of pixels, and after thresholding and filtering, are utilized to estimate a center of the colored object in the binary image.

2. The system of claim 1, wherein the data array includes pixels in a Hue-Saturation-Value (HSV) data array.

3. The system of claim 1, wherein every pixel in the binary image is represented by a “

0”

or a “

1”

, wherein each “

1”

represents a pixel that is determined to have similar charcteristics to those of the colored object.

4. The system of claim 1, wherein the first input source device is a digital camera.

5. The system of claim 1, wherein a search window is created in the binary image, the search window being centered around the estimated center of the colored object.

6. The system of claim 5, wherein a centroid of the “

1”

pixels in the search window is determined.

7. The system of claim 6, wherein the analysis system searches for pixels having predetermined characteristics located within a predetermined distance from the centroid.

8. An automated object tracking system for tracking a colored object through a series of frames of data, comprising:

a first image source device to provide a data array of pixels of a digital image;

a second image source device to provide a binary image of the data array; and

an analysis system to create a first histogram for columns of the binary image, and a second histogram for rows of the binary image, wherein each of the first histogram and the second histogram are thresholded and filtered, and after thresholding and filtering, are utilized to estimate a center of the colored object in the binary image,

wherein a search window is created in the binary image, the search window being centered around the estimated center of the colored object, a centroid of “

1”

pixels (“

1”

representing a pixel that is determined to have similar characteristics to those of the colored object) in the search window is determined, the analysis system searches for pixels having predetermined characteristics located within a predetermined distance from the centroid, and

if a predetermined number of pixels having predetermined characteristics are located within the predetermined distance from the centroid, the analysis system determines that the centroid is the center of the tracked object.

9. A method of automatically tracking a colored object through a series of frames of data, comprising:

providing a data array of pixels of a digital image, the digital image created by an image source;

providing a binary image of the data array; and

creating a first histogram for columns of the binary image and a second histogram for rows of the binary image,

thresholding and filtering the first histogram and the second histogram to create a first thresholded filtered histogram and a second thresholded filtered histogram, wherein the filtering reduces the effect of noise and reduces the effect of pixels in the data array that are away from a main concentration of pixels, and

utilizing the first thresholded filtered histogram and the second thresholded filtered histogram to estimate the center of the colored object in the binary image.

10. The method of claim 9, wherein the data array includes pixels in a Hue-Saturation-Value (HSV) data array.

11. The method of claim 9, wherein every pixel in the binary image is represented by a “

0”

or a “

1”

, wherein each “

1”

represents a pixel that is determined to have similar characteristics to those the colored object.

12. The method of claim 9, wherein the method further includes utilizing a digital camera as the image source.

13. The method of claim 9, wherein the method further includes creating a search window in the binary image, the search window being centered around the estimated center of the colored object.

14. The system of claim 13, wherein the method further includes determining a centroid of the “

1”

pixels in the search window.

15. The method of claim 14, wherein the method further includes searching for pixels having predetermined characteristics located within a predetermined distance from the centroid.

16. A method of automatically tracking a colored object through a series of frames of data, comprising:

providing a data array of pixels of a digital image;

providing a binary image of the data array;

creating a first histogram for columns of the binary image and a second histogram for rows of the binary image, wherein each histogram is thresholded and filtered, and after thresholding and filtering, are utilized to estimate the center of the colored object in the binary image;

creating a search window in the binary image, the search window being centered around the estimated center of the colored object;

determining a centroid of “

1”

pixels (“

1”

representing a pixel that is determined to have similar characteristics to those of the colored object) in the search window;

searching for pixels having predetermined characteristics located within a predetermined distance from the centroid; and

determining that the centroid is the center of the tracked object if a predetermined number of pixels having predetermined pixels are located within the predetermined distance from the centroid.

17. A method of automatically tracking a colored object through a series of frames of data, comprising:

providing a data array of pixels of a digital image;

creating a binary image from the data array of pixels by comparing each pixel in the data array of pixels to a pixel classification map wherein a first value represents that a pixel corresponds to an object in the pixel classification map and a second value represents that a pixel does not correspond to an object in the pixel classification map;

creating a first histogram for columns of the binary image and creating a second histogram for rows of the binary image;

filtering at least one of the first histogram and the second histogram to create a filtered first histogram and a filtered second histogram to reduce the effect of a remote column or remote row having a large number of pixels which have the first value that are away from a main concentration of pixels having the first value;

thresholding the filtered first histogram and the filtered second histogram to create a thresholded filtered first histogram and a thresholded filtered second histogram to erase columns and rows below a specified threshold level; and

utilizing the filtered thresholded first histogram and the filtered thresholded second histogram to estimate the center of the colored object in the binary image.

18. The method of claim 17, wherein filtering at least one of the first histogram or the second histogram to reduce the effect of a remote column or remote row having a large amount of pixels which have the first value that are away from a main concentration of pixels having the first value, includes analyzing a number of proximate columns or proximate rows in close proximity to the remote columns or remote rows.

19. The method of claim 18, further including as part of the filtering, calculating a new value of a number of pixels having the first value for the remote column or the remote row by adding a weighted value of the large amount of pixels of the remote column or the remote row to weighted values of a number of pixels in the number of proximate columns or proximate rows.

20. The method of claim 17, further including determining a center column in the filtered thresholded first histogram and a center row in the filtered thresholded second histogram, and creating a search window in the binary image, wherein the search window utilizes the center column and the center row as a search window center point.

21. The method of claim 20, wherein a number of columns in the search window is equal a number of pixels having the first value in the center column and a number of rows in the search window is equal to a number of pixels having the first value in the center row.

22. The method of claim 20, further including determining a centroid of the first value pixels in the search window.

23. The method of claim 22, further including searching for pixels having predetermined characteristics located within a predetermined distance from the centroid.

24. The method of claim 23, further including determining the centroid is the center of the tracked object if a predetermined number of pixels having predetermined characteristics are located within the predetermined distance from the centroid.