METHOD AND SYSTEM OF RAIL COMPONENT DETECTION USING VISION TECHNOLOGY

• US 20120263342A1
• Filed: 04/15/2011
• Published: 10/18/2012
• Est. Priority Date: 04/15/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Haas, Norman, Otto, Charles A., Pankanti, Sharathchandra, Li, Ying
• Granted Patent: US 8,625,878 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: B61L 23/047 Track or rail movements
• CPC Class Codes: B61L 23/048 Road bed changes, e.g. road bed erosion
• CPC Class Codes: G06F 18/00 Pattern recognition

Abstract

A method, system, and computer program product for automatically inspecting railroad tracks. The method includes assessing a configuration of rail components depicted in an image by comparing the configuration of the rail components to known hazards. The method also includes determining a severity of detected problems in the configuration of the rail components, using a computer processor.

Claims

1. A method for automatically inspecting railroad tracks comprising:

assessing a configuration of rail components depicted in an image by comparing the configuration of the rail components to safety requirements for the rail components; and

determining a severity of detected problems in the configuration of the rail components, using a computer processor.

2. The method of claim 1, wherein determining the severity of detected problems in the configuration is based on a discrepancy between an expected configuration model and a state of the configuration of the rail components.

3. The method of claim 2, wherein the expected configuration model is based on a surrounding.

4. The method of claim 2, wherein the expected configuration model is obtained from a database.

5. The method of claim 4 further comprising:

dividing the image into a set of cells;

extracting texture features from each cell from the set of cells;

grouping the set of cells into classes based on texture; and

binarizing the image.

6. The method of claim 5 further comprising:

detecting a tie plate from the image; and

detecting anchors on the left and right side of the tie plate.

7. The method of claim 5, wherein grouping the set of cells into classes based on texture is accomplished using at least one of Support Vector Machine (SVM), Gaussian Mixture Models (GMM), and K-means.

8. The method of claim 6 wherein detecting a tie plate from the image comprises:

finding the bottom line of a rail; and

identifying a region that contains a majority of pixels from a class of the classes of the set of cells, where the class has a greatest number of members.

9. The method of claim 6, wherein detecting anchors comprises at least one of template based matching, scale-invariant-feature-transform-based matching, and machine-learning based approaches.

10. The method of claim 4 further comprising:

detecting horizontal lines in the image;

interpreting the horizontal lines detected to determine a location of the rail and a vertical position of the tie plate; and

determining a horizontal boundary of the tie plate.

11. The method of claim 10, wherein detecting the horizontal lines comprises using the Hough transform.

12. The method of claim 10, wherein determining the horizontal boundary of the tie plate includes detecting vertical lines defining a left boundary and a right boundary of the tie plate using the Hough transform.

13. The method of claim 10, wherein determining the horizontal boundary of the tie plate includes analyzing the distribution of points which voted for the line at the bottom of the tie plate.

14. The method of claim 1, further comprising, separating foreground objects from background objects in the image, where foreground objects include at least one of a rail, a spike, a spike hole, a clip, a tie, a tie plate, and an anchor, and where background objects include at least one of a ballast object and a sky object.

15. The method of claim 1, further comprising, identifying other rail components in the image including spikes, spike holes, joint bars, joint bar bolts, nuts and washers.

16. The method of claim 1 further comprising, detecting the current condition of the rail components.

17. The method of claim 16 wherein detecting the current condition of the rail components comprises detecting the absence of rail components.

18. The method of claim 15 wherein identifying spikes comprises searching for any elliptical centers voted by edge points within a range of radius with an assumption of a fixed ratio of major axis to minor axis.

19. The method of claim 15 wherein identifying spike holes comprises searching for any regions-of-interest within an extended tie plate region that present strong edges with symmetric shapes.

20. A system for automatically inspecting railroad tracks comprising:

a processor;

a memory coupled to the processor, the memory having computer readable program code embodied therewith, the computer readable program code configured to;

assess a configuration of rail components located in an image by comparing the configuration of the rail components to safety requirements for the rail components; and

determine the severity of problems in the configuration of the rail components.

21. The system of claim 20 wherein the computer readable program code is further configured to link geographical location data with the image.

22. The system of claim 20, wherein the computer readable program code is further configured to determine the severity of problems in the configuration is based on a discrepancy between an expected configuration model and a state of the configuration of the rail components.

23. The method of claim 22, wherein the expected configuration model is obtained from a database.

24. A computer program product for automatically inspecting railroad tracks, the computer program product comprising:

a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code configured to;

assess a configuration of rail components located in an image by comparing the configuration of the rail components to safety requirements for the rail components; and

determine the severity of problems in the configuration of the rail components.

25. The computer program product of claim 24, wherein determining the severity of the problems in the configuration is based on a discrepancy between an expected configuration model and a state of the configuration of the rail components.