Visual profile classification

• US 20040197010A1
• Filed: 04/02/2003
• Published: 10/07/2004
• Est. Priority Date: 04/02/2003
• Status: Active Grant

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Inventors: Lee, Harry C., Olson, Teresa L., Sanders, Johnnie J.
• Granted Patent: US 6,901,152 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/421 by analysing segments intersecting the pattern

Abstract

An automated method of, and computer software and apparatus for, classifying objects visually into one of a plurality of object types comprising receiving a still image including an object, bounding the object within the image, dividing the bound portion of the image into a plurality of profile sections, performing a transform on each of the profile sections selected from discrete cosine transforms and discrete Fourier transforms, and executing a Bayes classifier function to segregate the object into one of the object types.

Claims

1. An automated method of classifying objects visually into one of a plurality of object types, the method comprising in the order presented the automated steps of:

receiving a still image including an object;

bounding the object within the image;

dividing the bound portion of the image into a plurality of profile sections;

performing a transform on each of the profile sections selected from the group consisting of discrete cosine transforms and discrete Fourier transforms; and

executing a Bayes classifier function to segregate the object into one of the object types.

2. The method of claim 1 wherein the object is a vehicle.

3. The method of claim 2 wherein the plurality of object types comprises a wheeled vehicle type and a tracked vehicle type.

4. The method of claim 1 wherein the executing step comprises executing a naive Bayes classifier function.

5. The method of claim 1 additionally comprising the step of performing clutter removal after the bounding step.

6. The method of claim 1 additionally comprising the step of performing a non-linear edge-enhancement function on the image before the dividing step.

7. The method of claim 6 additionally comprising the step of performing a Sobel function on the image before the dividing step.

8. The method of claim 7 additionally comprising the step of performing a thresholding function on the image after the step of performing a Sobel function and before the dividing step.

9. Computer software for classifying objects visually into one of a plurality of object types, said software comprising:

means for receiving a still image including an object;

means for bounding the object within the image;

means for dividing the bound portion of the image into a plurality of profile sections;

means for performing a transform on each of the profile sections selected from the group consisting of discrete cosine transforms and discrete Fourier transforms; and

means for executing a Bayes classifier function to segregate the object into one of the object types.

10. The software of claim 9 wherein said executing means comprises means for executing a naive Bayes classifier function.

11. The software of claim 9 additionally comprising means for performing clutter removal after bounding.

12. The software of claim 9 additionally comprising means for performing a non-linear edge-enhancement function on the image before dividing.

13. The software of claim 12 additionally comprising means for performing a Sobel function on the image before dividing.

14. The software of claim 13 additionally comprising means for performing a thresholding function on the image after performing a Sobel function and before dividing.

15. An apparatus for classifying objects visually into one of a plurality of object types, said apparatus comprising:

means for receiving a still image including an object;

means for bounding the object within the image;

means for dividing the bound portion of the image into a plurality of profile sections;

means for performing a transform on each of the profile sections selected from the group consisting of discrete cosine transforms and discrete Fourier transforms; and

means for executing a Bayes classifier function to segregate the object into one of the object types.

16. The apparatus of claim 15 wherein said executing means comprises means for executing a naive Bayes classifier function.

17. The apparatus of claim 15 additionally comprising means for performing clutter removal after bounding.

18. The apparatus of claim 15 additionally comprising means for performing a non-linear edge-enhancement function on the image before dividing.

19. The apparatus of claim 18 additionally comprising means for performing a Sobel function on the image before dividing.

20. The apparatus of claim 19 additionally comprising means for performing a thresholding function on the image after performing a Sobel function and before dividing.