Stationary target detection by exploiting changes in background model

• US 8,526,678 B2
• Filed: 02/19/2013
• Issued: 09/03/2013
• Est. Priority Date: 09/04/2007
• Status: Active Grant

• Current Assignee: Motorola Solutions, Inc.
• Original Assignee: ObjectVideo Incorporated (Alarm.com Holdings, Inc.)
• Inventors: Liu, Haiying, Yu, Li, Hassan-Shafique, Khurram, Venetianer, Pter L.
• Primary Examiner(s): Kassa, Yosef
• Application Number:
  US13/770,700
• Publication Number:
  US 20130188833A1
• Time in Patent Office: 196 Days
• Field of Search: 382/103, 382/170, 382/171, 382/209, 382/278, 382/282, 358/538, 358/464
• US Class Current: 382/103
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 18/24 Classification techniques
• 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

Abstract

One or more video frames may be obtained, and a background model may be constructed based on a first parameter. A second background model may be constructed using the one or more video frames based on a second parameter, the second parameter being different from the first parameter. A difference between the first and second background models may be determined. One or more stationary targets may be determined based on the determined difference. The one or more stationary targets may be classified. An alert concerning the one or more classified stationary targets may be generated.

Claims

1. A computer-implemented method for processing one or more video frames, comprising:

obtaining, by a computer, one or more video frames;

constructing, by the computer, a first background model using the one or more video frames based on a first parameter;

constructing, by the computer, a second background model using the one or more video frames based on a second parameter, the second parameter being different from the first parameter;

determining, by the computer, a difference between the first and second background models;

determining, by the computer, one or more stationary targets based on the determined difference;

classifying, by the computer, the one or more stationary targets to obtain one or more classified stationary targets; and

generating, by the computer, one or more alerts concerning the one or more classified stationary targets.

2. A computer-implemented method for processing one or more video frames, comprising:

obtaining, by a computer, one or more video frames;

constructing, by the computer, a first background model using the one or more video frames based on a first parameter;

constructing, by the computer, a second background model using the one or more video frames based on a second parameter, the second parameter being different from the first parameter;

determining, by the computer, a difference between the first background model and the second background model;

determining, by the computer, at least one of one or more insertion targets or one or more removal targets.

3. The method as set forth in claim 2, further comprising:

providing an alert for at least one of the one or more insertion targets or the one or more removal targets.

4. The method as set forth in claim 2, wherein determining at least one of the one or more insertion targets or the one or more removal targets comprises:

generating one or more blobs using the one or more video frames;

computing one or more first gradients within the one or more blobs for the first background model;

computing one or more second gradients within the one or more blobs for the second background model;

computing one or more correlations of the one or more first gradients and the one or more second gradients within the one or more blobs;

comparing the computed one or more correlations to a first threshold; and

classifying the one or more blobs as at least one of the one or more insertion targets or the one or more removal targets.

5. The method as set forth in claim 2, further comprising:

determining one or more blobs as one or more local changes based on the computed one or more correlations being greater than the first threshold.

6. The method as set forth in claim 2, wherein determining at least the one or more insertion targets or the one or more removal targets comprises:

computing first gradient strengths of the one or more first gradients at one or more boundaries of the one or more blobs for the first background model;

computing second gradient strengths of the one or more second gradients at one or more boundaries of the one or more blobs for the second background model;

determining one or more ratios of the first gradient strengths to the second gradient strengths;

comparing the one or more ratios to a second threshold; and

classifying the one or more blobs as at least one of the one or more target insertions or the one or more target removals based on the comparison of the one or more ratios to the second threshold.

7. A computer-implemented method for processing one or more video frames, comprising:

obtaining, by a computer, one or more video frames;

generating, by the computer, one or more blobs using the one or more video frames;

classifying, by the computer, the one or more blobs to produce one or more classified blobs, wherein the one or more classified blobs include one or more target insertions; and

determining, by the computer, one or more stationary targets based on a length of time of insertion for the one or more target insertions.

8. The method of claim 7, wherein determining one or more stationary targets further comprises:

constructing a list of detected targets based on the classified blobs;

comparing one or more blobs with the list of detected targets to detect one or more target matches;

determining one or more number of slices for the one or more target matches;

detecting one or more stationary targets based on comparing the one or more number of slices to one or more stationary thresholds; and

classifying the one or more blobs as one or more stationary targets based on the comparing to the one or more stationary thresholds.

9. The method of claim 8, wherein the one or more number of slices is determined based on counting a number of frames where the one or more target matches are detected in the one or more video frames.

10. The method of claim 8, further comprising:

generating one or more alerts regarding the one or more determined stationary targets.

11. The method of claim 8, further comprising:

adding a blob from the one or more blobs to the list of targets when the blob is not in the list of detected targets.

12. The method of claim 8, further comprising:

classifying one or more blobs in the list of targets as one or more local changes when the one or more number of slices does not meet the one or more stationary thresholds; and

removing the one or more blobs classified as one or more local changes from the list of targets.