Target detection and tracking from overhead video streams

• US 7,796,780 B2
• Filed: 06/24/2005
• Issued: 09/14/2010
• Est. Priority Date: 06/24/2005
• Status: Active Grant

• Current Assignee: Motorola Solutions, Inc.
• Original Assignee: ObjectVideo Incorporated (Alarm.com Holdings, Inc.)
• Inventors: Gupta, Himaanshu, Yu, Li, Zhang, Zhong, Venetianer, Peter L., Lipton, Alan J., Liu, Haiying, Rasheed, Zeeshan
• Primary Examiner(s): Carter; Aaron W
• Application Number:
  US11/165,435
• Publication Number:
  US 20060291695A1
• Time in Patent Office: 1,908 Days
• Field of Search: 382/103, 382/104, 382/197, 382/228, 382/107, 345/419, 348143-160, 348169-172
• US Class Current: 382/103
• CPC Class Codes: G06T 7/215 Motion-based segmentation
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• 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 13/19608 Tracking movement of a target, e.g. by detecting an object predefined as a target, using target direction and or velocity to predict its new position

Abstract

A technique for video processing includes: receiving video from an overhead view of a scene; detecting moving pixels in the video; detecting line segments in the video based on detected moving pixels; identifying targets in the video based on the detected line segments; tracking targets in the video based on the identified targets; and managing tracked targets in the video.

Claims

1. A non-transitory computer-readable medium comprising software for video processing, which when executed by a computer system, cause the computer system to perform operations comprising a method of:

receiving video from an overhead view of a scene;

detecting moving pixels in the video;

detecting line segments in a foreground of the scene in the video based on detected moving pixels;

identifying targets in the video based on the detected line segments;

tracking targets in the video based on the identified targets; and

managing tracked targets in the video, wherein detecting line segments comprises;

counting edge pixels; and

identifying a line segment based on the edge pixels.

2. A computer-readable medium as in claim 1, wherein detecting moving pixels comprises:

separating the foreground in the video from background in the video; and

detecting foreground edges in the video.

3. A computer-readable medium as in claim 1, wherein identifying the line segment comprises:

identifying a start point;

predicting next search directions;

identifying a next line pixel; and

providing a line segment.

4. A computer-readable medium as in claim 1, wherein identifying targets comprises:

updating existing targets;

detecting new targets;

refining the new targets;

merging the existing targets and the new targets;

splitting the existing targets and the new targets; and

cleaning the existing targets and the new targets.

5. A computer-readable medium as in claim 4, wherein updating targets comprises:

predicting a target on its new location;

assigning detected line segments to the predicted target; and

updating the target.

6. A computer-readable medium as in claim 4, wherein detecting new targets comprises:

performing line segment clustering;

performing cluster verification based on the line segment clustering; and

generating a new target based on the cluster verification.

7. A computer-readable medium as in claim 4, wherein refining the new targets comprises:

agglomerating remaining line segments to nearest targets;

re-estimating the targets; and

updating the targets.

8. A computer-readable medium as in claim 4, wherein merging the existing targets and the new targets comprises:

obtaining a target pair; and

merging the target pair if parameters of the target pair are similar.

9. A computer-readable medium as in claim 4, wherein splitting the existing targets and the new targets comprises:

obtaining a target;

performing line clustering on the obtained target if the obtained target is not similar to a normal target to obtain clusters; and

assigning a target identities to the clusters if the clusters are similar to the normal target.

10. A computer-readable medium as in claim 4, wherein merging and splitting the existing targets and the new targets comprises:

generating a foreground mask;

detecting foreground objects based on the foreground mask; and

analyzing the foreground objects to obtain a number of targets.

11. A computer-readable medium as in claim 10, wherein analyzing the foreground objects is based on comparing the foreground objects to a multiple target size threshold, a minimum single target size threshold, and a maximum single target size threshold.

12. A computer-readable medium as in claim 4, wherein cleaning the existing targets and the new targets comprises:

obtaining a target;

maintaining the obtained target if the obtained target is detected in a current frame or was not moving out of a field of view of the video camera; and

removing the obtained target if the obtained target is not detected in a current frame and was moving out of a field of view of the video camera.