Non-motion detection

• US 7,688,997 B2
• Filed: 11/21/2002
• Issued: 03/30/2010
• Est. Priority Date: 11/21/2001
• Status: Active Grant

• Current Assignee: iOmniscient Pty Ltd.
• Original Assignee: iOmniscient Pty Ltd.
• Inventors: Gibbins, Daniel, Brooks, Michael John, Newsam, Garry
• Primary Examiner(s): Carter; Aaron W
• Application Number:
  US10/496,424
• Publication Number:
  US 20050036658A1
• Time in Patent Office: 2,686 Days
• Field of Search: 382/103, 382/107, 382/218, 348143-160
• US Class Current: 382/103
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G08B 13/19602 Image analysis to detect motion of the intruder, e.g. by frame subtraction
• CPC Class Codes: G08B 13/19604 involving reference image or background adaptation with time to compensate for changing conditions, e.g. reference image update on detection of light level change
• CPC Class Codes: G08B 13/19606 Discriminating between target movement or movement in an area of interest and other non-signicative movements, e.g. target movements induced by camera shake or movements of pets, falling leaves, rotating fan
• CPC Class Codes: G08B 13/19682 Graphic User Interface [GUI] presenting system data to the user, e.g. information on a screen helping a user interacting with an alarm system
• CPC Class Codes: G08B 13/19693 using multiple video sources viewed on a single or compound screen

Abstract

An image sequence derived from area surveillance is considered as a collection of independent time series, each of which describes an image feature that summarizes a small neighborhood of the image. These time series are regarded as samples from some underlying distribution, and statistical techniques are used to identify the density function of this distribution. The background can then be identified as a feature of this distribution, moreover the distribution of features in subsequent image frames can be monitored to see if it deviates sufficiently from the current distributions to indicate a change in the background.

Claims

1. A method of detecting a change to a background of a visual scene, the method comprising, for a current image of the visual scene:

extracting a feature value from at least one sub-region of the current image;

updating a detection distribution using the extracted feature value;

determining whether a match exists between the updated detection distribution and a background distribution;

comparing with a predetermined threshold, if a match is not determined, the number of images since the previously determined match; and

detecting, based on the comparison, a change to the background of the visual scene,

wherein the extracting, updating, determining, comparing, and detecting are performed by a processor.

2. The method according to claim 1, further comprising:

resetting, if a match is determined, the detection distribution.

3. The method according to claim 1, wherein the determining comprises:

determining whether a most prominent feature value in the background distribution is sufficiently prominent in the updated detection distribution.

4. The method according to claim 1, wherein the detecting comprises:

flagging, based on the comparison, a possible change to the background of the visual scene;

comparing, based on the flagging, the updated detection distribution with a predetermined limit;

detecting, based on the further comparison, a change to the background of the visual scene.

5. The method according to claim 4, further comprising:

determining whether the number of images since the previously determined match exceeds a second threshold; and

detecting, based on the determination, a change to the background of the visual scene.

6. The method according to claim 1, further comprising:

detecting whether the sub-region contains motion; and

carrying out the updating, determining, comparing, and detecting dependent on whether the sub-region contains motion.

7. The method according to claim 6, wherein the detecting of motion comprises determining whether the extracted feature is substantially different to the feature extracted from the previous image.

8. The method according to claim 1, wherein the extracted feature is a regularized pattern feature.

9. The method according to claim 1, further comprising updating the background distribution using the extracted feature.

10. The method according to claim 1, further comprising comparing the shape of a plurality of adjacent sub-regions for which a change to the background has been detected with a shape template.

11. The method according to claim 1, wherein the background distribution and the detection distribution are histograms of extracted feature values.

14. The method according to claim 1, wherein the comparing comprises:

incrementing a counter; and

comparing the counter with the predetermined threshold.

15. The method according to claim 14, further comprising resetting, if a match is determined, the counter to zero.

12. A computer readable storage medium, having recorded thereon computer software code for performing a method of detecting a change to a background of a visual scene, comprising, for a current image of the visual scene:

extracting a feature value from at least one sub-region of the current image;

updating a detection distribution using the extracted feature value;

determining whether a match exists between the updated detection distribution and a background distribution;

comparing with a predetermined threshold, if a match is not determined, the number of images since the previously determined match; and

detecting, based on the comparison, a change to the background of the visual scene.

13. A computer system for detecting a change to a background of a visual scene, the computer system comprising:

an input port for accepting images of the scene into which objects enter and from which objects leave; and

process means for, for a current image of the visual scene;

extracting a feature value from at least one sub-region of the current image;

updating a detection distribution using the extracted feature value;

determining whether a match exists between the updated detection distribution and a background distribution;

comparing with a predetermined threshold, if a match is not determined, the number of images since the previously determined match; and

detecting, based on the comparison, a change to the background of the visual scene.