Assisted surveillance of vehicles-of-interest

• US 9,864,923 B2
• Filed: 06/20/2016
• Issued: 01/09/2018
• Est. Priority Date: 03/13/2012
• Status: Active Grant

• Current Assignee: Massachusetts Institute of Technology
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Chan, Michael T., Thornton, Jason R., Yahr, Aaron Z., Zwahlen, Heather
• Primary Examiner(s): EL CHANTI, HUSSEIN A
• Application Number:
  US15/187,347
• Publication Number:
  US 20160321519A1
• Time in Patent Office: 568 Days
• CPC Class Codes: G06F 2218/12 Classification; Matching
• 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: G06V 20/54 of traffic, e.g. cars on the road, trains or boats
• CPC Class Codes: G06V 2201/08 Detecting or categorising vehicles
• CPC Class Codes: G08G 1/015 with provision for distinguishing between two or more types of vehicles, e.g. between motor-cars and cycles

Abstract

A computer-implemented image processing method includes accessing image data representing a plurality of images of a plurality of vehicles and reading a data model representing a predefined vehicle attribute. The method further includes processing a portion of the image data to detect a vehicle feature of at least one of the vehicles represented in at least one of the images based on the predefined vehicle attribute, and processing the portion of the image data to generate vehicle data representing the detected vehicle feature.

Claims

1. A computer-implemented image processing method, the computer including a processor and a memory operatively coupled to the processor, the method performed by the processor comprising:

accessing image data representing a plurality of images of a plurality of vehicles;

accessing a discriminative type-specific vehicle model calculated from training data representing a predefined vehicle attribute;

processing, using the discriminative type-specific vehicle model, a portion of the image data to automatically detect a vehicle feature of a vehicle of interest from the plurality of vehicles represented in two or more images of the plurality of images based on the predefined vehicle attribute;

processing the portion of the image data to automatically generate vehicle data representing the detected vehicle feature; and

reconstructing a path traveled by the vehicle of interest using the vehicle data.

2. The computer-implemented method of claim 1, wherein the discriminative type-specific vehicle model is a part-based mixture model calculated from training data representing the predefined vehicle attribute and where the vehicle feature is detected in the portion of the image data using the part-based mixture model.

3. The computer-implemented method of claim 2, wherein the part-based mixture model is learned using a latent support vector machine.

4. The computer-implemented method of claim 1, further comprising processing the image data representing the detected vehicle feature to generate a characterization of the vehicle of interest based at least in part on the discriminative type-specific vehicle model.

5. The computer-implemented method of claim 4, wherein the characterization of the vehicle of interest includes at least one of a predefined vehicle type and a predefined vehicle color.

6. The computer-implemented method of claim 1, further comprising:

receiving a search query from a user;

processing the vehicle data to search for the vehicle of interest in the plurality of images corresponding to the search query; and

displaying, using a graphical user interface operatively coupled to the processor, a composite view depicting recorded activities of the vehicle of interest across multiple camera views.

7. The computer-implemented method of claim 1, further comprising exporting a composite image depicting recorded activities of the vehicle of interest across multiple camera views.

8. An image processing system comprising:

a processor; and

a memory operatively coupled to the processor, the memory having stored therein instructions that when executed by the processor cause the processor to;

access image data representing a plurality of images of a plurality of vehicles;

access a discriminative type-specific vehicle model calculated from training data representing a predefined vehicle attribute;

process, using the discriminative type-specific vehicle model, a portion of the image data to automatically detect a vehicle feature of a vehicle of interest from the plurality of vehicles represented in two or more images of the plurality of images based on the predefined vehicle attribute;

process the portion of the image data to automatically generate vehicle data representing the detected vehicle feature; and

reconstruct a path traveled by the vehicle of interest using the vehicle data.

9. The system of claim 8, wherein the memory has further stored therein instructions that when executed by the processor cause the processor to:

receive a search query from a user;

process the vehicle data to search for the vehicle of interest in the plurality of images corresponding to the search query; and

display, using a graphical user interface operatively coupled to the processor, a composite view depicting recorded activities of the vehicle of interest across multiple camera views.

10. The system of claim 8, wherein the memory has further stored therein instructions that when executed by the processor cause the processor to export a composite image depicting recorded activities of the vehicle of interest across multiple camera views.

11. The system of claim 8, wherein the memory has further stored therein instructions that when executed by the processor cause the processor to process the image data representing the detected vehicle feature to generate a characterization of the vehicle of interest based at least in part on the discriminative type-specific vehicle model.

12. A computer-implemented image processing method, the computer including a processor and a memory operatively coupled to the processor, the method performed by the processor comprising:

accessing image data representing a plurality of images of a plurality of vehicles;

accessing a discriminative type-specific vehicle model calculated from training data representing a predefined vehicle attribute;

processing, using the discriminative type-specific vehicle model, a portion of the image data to automatically detect a vehicle feature of a vehicle of interest from the plurality of vehicles represented in two or more images of the plurality of images based on the predefined vehicle attribute;

processing the portion of the image data to automatically generate vehicle data representing the detected vehicle feature; and

determining a direction of motion of the vehicle of interest using the vehicle data.

13. The computer-implemented method of claim 12, further comprising exporting a composite image depicting recorded activities of the vehicle of interest across multiple camera views.

14. The computer-implemented method of claim 12, wherein the discriminative type-specific vehicle model is a part-based mixture model calculated from training data representing the predefined vehicle attribute and where the vehicle feature is detected in the portion of the image data using the part-based mixture model.

15. The computer-implemented method of claim 14, wherein the part-based mixture model is learned using a latent support vector machine.

16. The computer-implemented method of claim 12, further comprising processing the image data representing the detected vehicle feature to generate a characterization of the vehicle of interest based at least in part on the discriminative type-specific vehicle model.

17. The computer-implemented method of claim 16, wherein the characterization of the vehicle includes at least one of a predefined vehicle type and a predefined vehicle color.

18. The computer-implemented method of claim 12, further comprising:

receiving a search query from a user;

processing the vehicle data to search for a vehicle of interest in the plurality of images corresponding to the search query; and

displaying, using a graphical user interface operatively coupled to the processor, a composite view depicting recorded activities of the vehicle of interest across multiple camera views.