Video-tracking for video-based speed enforcement

• US 8,971,573 B2
• Filed: 09/12/2012
• Issued: 03/03/2015
• Est. Priority Date: 09/12/2012
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Wu, Wencheng
• Primary Examiner(s): Lefkowitz, Sumati
• Assistant Examiner(s): JAMES, THOMAS AIDEN
• Application Number:
  US13/611,718
• Publication Number:
  US 20140072168A1
• Time in Patent Office: 902 Days
• Field of Search: None
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/30236 Traffic on road, railway or crossing
• CPC Class Codes: G06T 7/248 involving reference images or patches
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

A method for tracking a moving vehicle includes detecting the vehicle by acquiring a series of temporal related image frames. In an initial frame, the detecting includes locating a reference feature representative of a vehicle. The method further includes setting the reference feature as a full-size template. The method includes tracking the vehicle by searching a current frame for features matching the full-size template and at least one scaled template. The tracking further includes setting as an updated template one of the full-size and scaled templates closest matching the feature. The method includes repeating the tracking using the updated template for each next frame in the series.

Claims

1. A method for tracking a moving vehicle, the method comprising:

detecting a vehicle, the detecting including;

acquiring a series of temporal related image frames, and

in an initial frame, locating a reference feature representative of a vehicle;

setting the reference feature as a full-size template;

tracking the vehicle, the tracking including;

searching a current frame for a feature matching the full-size template,

searching the current frame for a feature matching at least one scaled template,

setting as an updated template one of the full-size and scaled templates closest matching the feature; and

repeating the tracking using the updated template for each next frame in the series.

2. The method of claim 1 further comprising:

using the tracking, computing a distance that the feature moves between select frames in the series using a location of the feature in the select frames; and

,

computing a speed using the distance.

3. The method of claim 1, wherein the feature includes at least one portion of a license plate.

4. The method of claim 1 further comprising:

scaling the full-size template by a predetermined amount to generate the scaled template.

5. The method of claim 4, wherein the predetermined scaling amount is dynamic amount using at least one of camera calibration data, a current location of the feature, an average vehicle speed of a current traffic scene being monitored, and a scaling value from a previously tracked feature.

6. The method of claim 1, wherein the searching the current frame for a feature matching the full size template includes:

computing a first correlation number for describing a similarity between the feature and the full-size template; and

,

storing the first correlation number and a location of the feature in the current frame;

wherein the searching the current frame for a feature matching the at least one scaled template includes;

computing a second correlation number for describing a similarity between feature and the scaled template; and

,

storing the second correlation number and the location of the feature in the current frame.

7. The method of claim 6, wherein the setting the updated template includes:

comparing the first and second correlation numbers;

determining a higher one of the first and second correlation numbers;

setting as the updated template the one of the full-size template and scaled template corresponding to the higher correlation number.

8. The method of claim 1 further comprising:

cropping a region around the feature; and

,

setting the cropped region as the full size template.

9. A computer program product comprising non-transitory computer-readable recording media which encodes instructions for performing the method of claim 1.

10. A system for tracking a moving vehicle, the system comprising:

a tracking device comprising memory which stores instructions for performing the method of claim 1 and a processor, in communication with the memory for executing the instructions.

11. A system for tracking a moving vehicle, the system comprising:

a vehicle tracking device, including;

an image capture module adapted to acquire a series of temporal related image frames

a vehicle detection module adapted to;

in an initial frame, locate a reference feature representative of a moving object;

a vehicle tracking module adapted to;

set the reference feature as a full-size template,

search a current frame for a feature matching the full-size template,

search the current frame for a feature matching at least one scaled template,

set as an updated template one of the full-size and scaled templates closest matching the feature, and

repeat the tracking using the updated template for each next frame in the series; and

,

a processor adapted to implement the modules.

12. The system of claim 11 further comprising a speed computation module adapted to:

using the tracking, compute a distance that the feature moves between a first frame and a last frame that the vehicle is detected in the series;

compute a speed using the distance.

13. The system of claim 11, wherein the feature includes at least one portion of a license plate.

14. The system of claim 11, wherein the vehicle tracking module is further adapted to:

scale the full-size template a predetermined amount to generate the scaled template.

15. The system of claim 14, wherein the predetermined amount is a dynamic amount being computed using one of camera calibration data, a current location of the feature, an average vehicle speed of a current traffic scene being monitored, and a scaling value from a previously tracked feature.

16. The system of claim 12, wherein the vehicle tracking module is further adapted to:

compute a first correlation number for describing a similarity between the feature and the full-size template; and

,

compute a second correlation number for describing a similarity between feature and the scaled template; and

,

store the first and second correlation numbers with a location of the feature in the current frame.

17. The system of claim 16, wherein the vehicle tracking module is further adapted to:

compare the first and second correlation numbers;

determine a higher one of the first and second correlation numbers;

set as the updated template the one of the full-size template and scaled template corresponding to the higher correlation number.

18. The system of claim 12, wherein the vehicle detection module is further adapted to:

crop a region around the feature; and

,

set the cropped region as the full size template.

19. The system of claim 12 further comprising a video capture device adapted to capture the temporal related image frames and transmit the frames to the vehicle detection module.

20. A method for tracking a moving vehicle, the method comprising:

detecting a vehicle, the detecting including;

acquiring a series of temporal related image frames, and

in an initial frame, locating a reference feature representative of a moving object;

setting the reference feature as a full-size template;

scaling the full-size template predetermined amounts to generate multiple scaled templates;

tracking the moving object, the tracking including;

searching the current frame for a feature matching the scaled templates,

computing a correlation number for describing a similarity between the detected feature and each of the full-size template and scaled templates;

setting as an updated template one of the full-size template and scaled templates corresponding to highest correlation number; and

storing a location of the feature in the current frame associated with the highest correlation number; and

,

repeating the tracking for each next frame in the series.