Digital image edge detection and road network tracking method and system

• US 20030223615A1
• Filed: 06/04/2002
• Published: 12/04/2003
• Est. Priority Date: 06/04/2002
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Jiang, Qin, Porikli, Fatih, Keaton, Patricia A.
• Granted Patent: US 7,636,455 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: G06T 17/05 Geographic models
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30256 Lane; Road marking
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06V 10/443 by matching or filtering
• CPC Class Codes: G06V 20/13 Satellite images

Abstract

A system and method for finding edge orientation and magnitude, and extracting a road path from an image. An image is segmented into at least a first window and a second window. A first road segment having plural parameters within the first window is specified. Plural road segments are identified, each having plural parameters, in the second window. The orientation of plural line segments in the second window are determined, and paired, based on their orientation. The line segment orientations are determined by receiving a plurality of edge orientation estimates, transforming a portion of them, and aggregating the transformed portion into an edge orientation estimate. A stochastic process is performed on the plural parameters of the plural road segments to identify a second road segment having the maximum correlation with the plural parameters of the first road segment.

Claims

1. A system for determining an edge orientation in an image, comprising:

means for receiving a plurality of edge orientations estimates;

means for transforming a portion of said plurality of edge estimates, and

a means for aggregating said transformed portion of said plurality of edge estimates.

2. The system of claim 1 wherein said image is a color image or a multi-spectral image.

3. The system of claim 1 wherein said edge orientation estimates are received from a plurality of directional filters, a plurality of compass filters, a plurality of channels, or plural outputs of a finite impulse response filter.

4. The system of claim 1 wherein said means for transforming utilizes a line to vector transform processor.

5. The system of claim 1 wherein said means for transforming edge responses such that the plurality of edge estimates is mapped from the 1-½

D direction domain to the 2D vector domain.

6. The system of claim 1 further comprising a means for receiving a magnitude associated with each edge orientation estimate and wherein said means for transforming is operable to convert scalar edge orientation and magnitude values to vectors.

7. The system of claim 1 wherein said means for aggregating comprises a means for fusing said transformed portions of said plurality of edge estimates.

8. The system of claim 1 where said means for aggregating applies a fusing algorithm to aggregate said transformed portions of said plurality of edge estimates.

9. The system of claim 1 wherein said means for aggregating fuses said plurality of transformed portions of said plurality of edge estimates using linear operations.

10. The system of claim 1 wherein said edge estimate comprises an orientation and a magnitude of the edge.

11. A system of extracting a road path from an image, comprising:

means for segmenting an image into at least a first window and a second window;

means for specifying a first road segment having plural parameters within said first window;

means for identifying plural road segments, each having plural parameters, in said second window; and

means for performing a stochastic process on said plural parameters of said plural road segments to identify a second road segment having the maximum correlation with said plural parameters of said first road segment.

12. The system of claim 11 wherein said means for segmenting selects groups of adjacent pixels in a digital image.

13. The system of claim 11 wherein said first window and said second window are adjacent to one another.

14. The system of claim 11 wherein said means for identifying is user input to the system.

15. The system of claim 11 wherein said means for identifying is a computer that operates automatically.

16. The system of claim 11 wherein said means for identifying further comprises a means for determining the orientation of plural line segments in said second window and a means for pairing said plural line segments based on a their orientation.

17. The system of claim 16 wherein said means for determining further comprises:

means for receiving a plurality of edge orientations estimates;

means for transforming a portion of said plurality of edge estimates; and

means for aggregating said transformed portion of said plurality of edge estimates.

18. The system of claim 11 wherein said plural parameters are comprised of road width, road orientation, road intensity, or road color.

19. The system of claim 11 wherein said stochastic process includes probabilistic modeling.

20. The system of claim 11 wherein said stochastic process includes a Markovian process to maximize the joint probability of said parameters.

21. The system of claim 11 wherein said parameters are processed in a said stochastic process as random variables to maximize the joint probability of said parameters.

22. The system of claim 11 further comprising a means for connecting said first road segment and said second road segment into a road path.

23. The system of claim 22 further comprising a non-linear filter that operates to reduce road path estimation errors.

24. A method of determining an edge orientation in an image, comprising the steps of:

receiving a plurality of edge orientations estimates;

transforming a portion of said plurality of edge estimates; and

aggregating said transformed portion of said plurality of edge estimates.

25. The method of claim 24 wherein said image is a color image or a multi-spectral image.

26. The method of claim 24 wherein said edge orientation estimates are received from a plurality of directional filters, a plurality of compass filters, a plurality of channels, or plural outputs of a finite impulse response filter.

27. The method of claim 24 wherein said transforming step utilizes a line to vector transform.

28. The method of claim 24 wherein said transforming step maps said plurality of edge estimates from the 1-½

D directional to the 2D vector domains.

29. The method of claim 24 further comprising the step of receiving a magnitude associated with each edge orientation estimate and wherein said transforming step operates to convert scalar edge orientation and magnitude values to vectors.

30. The method of claim 24 wherein said aggregating step comprises the step of fusing said transformed portions of said plurality of edge estimates.

31. The method of claim 24 where said aggregating step applies a fusing algorithm to aggregate said transformed portions of said plurality of edge estimates.

32. The method of claim 24 wherein said aggregating step fuses said plurality of transformed portions of said plurality of edge estimates using linear operations.

33. The method of claim 24 wherein said edge estimate comprises an orientation and a magnitude of the edge.

34. A method of extracting a road path from an image, comprising the steps of:

segmenting an image into at least a first window and a second window;

specifying a first road segment having plural parameters within said first window;

identifying plural road segments, each having plural parameters, in said second window; and

performing a stochastic process on said plural parameters of said plural road segments to identify a second road segment having the maximum correlation with said plural parameters of said first road segment.

35. The method of claim 34 wherein said segmenting step is accomplished by selecting groups of adjacent pixels in a digital image.

36. The method of claim 34 wherein said first window and said second window are adjacent to one another.

37. The method of claim 34 wherein said specifying step is accomplished by manual user intervention.

38. The method of claim 34 wherein said specifying step is preformed automatically by a computer.

39. The method of claim 34 wherein said identifying step further comprises the steps of determining the orientation of plural line segments in said second window and pairing said plural line segments based on a their orientation.

40. The method of claim 39 wherein said determining step further comprises the steps of:

receiving a plurality of edge orientations estimates;

transforming a portion of said plurality of edge estimates; and

aggregating said transformed portion of said plurality of edge estimates.

41. The method of claim 34 wherein said plural parameters are comprised of road width, road orientation, road intensity, or road color.

42. The method of claim 34 wherein said stochastic process includes probabilistic modeling.

43. The method of claim 34 wherein said stochastic process includes a Markovian process to maximize the joint probability of said parameters.

44. The method of claim 34 wherein said parameters are processed in a said stochastic process as random variables to maximize the joint probability of said parameters.

45. The method of claim 34 further comprising the step of connecting said first road segment and said second road segment into a road path.

46. The method of claim 45 further comprising the step of employing a non-linear filter to reduce road path estimation errors.