Method and apparatus for the automated detection and classification of defects in sewer pipes

• US 7,215,811 B2
• Filed: 11/23/2001
• Issued: 05/08/2007
• Est. Priority Date: 11/22/2000
• Status: Active Grant

• Current Assignee: Osama Moselhi, Tariq Shebab-Eldeen
• Original Assignee: Osama Moselhi, Tariq Shebab-Eldeen
• Inventors: Moselhi, Osama, Shehab-Eldeen, Tariq
• Primary Examiner(s): CHAWAN, SHEELA C
• Application Number:
  US09/990,572
• Publication Number:
  US 20030023404A1
• Time in Patent Office: 1,992 Days
• Field of Search: 382/100, 382/156, 382/157, 382/190, 382/191, 382/192, 382/195, 382/218, 382/224, 382/228, 382/276, 382/173, 382/149, 735/87, 735/92, 736/23, 736/18, 736/19, 348/84, 348/88, 348/125, 348/91, 348/92, 348/128, 707/101, 707/100, 707/200, 707/6, 707/2, 706/1, 706/15, 706/25, 376/250, 376/251, 376/252, 324219-221, 324/512, 324/753
• US Class Current: 382/156
• CPC Class Codes: G01N 21/954 Inspecting the inner surface of hollow bodies, e.g. bores
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20036 Morphological image processing
• CPC Class Codes: G06T 2207/30132 Masonry; Concrete
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume

Abstract

In accordance with a first aspect of the present invention, there is provided a method for detecting a defect on a portion of an element comprising the steps of: acquiring an image of said portion; analyzing said image to highlight problematic regions of said portion; calculating a probability that said problematic region is a defect; if said probability is higher than a threshold value, determining a position of said defect on said element. Another method for classifying a defect on an element is provided. The method comprises: acquiring an image of said defect; calculating a probability that said defect corresponds to one of a series of types of defects; if said probability is higher than a threshold value, determining that said defect is a defect of that particular type. Another method for recommending a most suitable rehabilitation technique for a defect is provided. The method comprises: identifying a series of parameters corresponding to said defect; calculating a relative utility for each of a series of potential rehabilitation techniques using rehabilitation profiles; determining a most suitable rehabilitation technique for said defect corresponding to a highest value of said relative utility.

Claims

1. A method for detecting a plurality of defects in an item under inspection comprising:

acquiring at least one image of said item;

providing a plurality of neural networks, at least one of said plurality of neural networks corresponding to each one of said plurality of defects to be detected, wherein each one of said plurality of defects is selected from at least one of cross-sectional reductions, misalignments, infiltration, and cracks;

processing said at least one image to produce a processed image having objects isolated from an image background of said image; and

inputting said processed image into said plurality of neural networks to obtain information concerning corresponding defects.

2. A method as claimed in claim 1, further comprising issuing a report using said information concerning said defects.

3. A method as claimed in claim 2, further comprising recommending a rehabilitation technique based on said report and a set of attributes of said item under inspection.

4. A method as claimed in claim 3, wherein said attributes are part of a group comprising technical requirements, contractual requirements, and cost effectiveness.

5. A method as claimed in claim 3, wherein a plurality of rehabilitation techniques are recommended.

6. A method as claimed in claim 5, further comprising ranking said plurality of recommended rehabilitation techniques.

7. A method as claimed in claim 1, wherein said plurality of neural networks further comprises sets of neural networks used for counter-checking results, each one of said sets of similar neural networks corresponding to each one of said plurality of defects to be detected.

8. A method as claimed in claim 1, wherein each set of neural networks comprises at least three neural networks used for counter-checking results.

9. A method as claimed in claim 1, wherein processing said at least one image further comprises processing said at least one image according to a selected set of image analysis techniques, said set of image analysis techniques selected as a function of said defects to be detected.

10. A method as claimed in claim 9, wherein n sets of neural networks are used to detect n types of defects.

11. A method as claimed in claim 10, wherein:

said item under inspection is a sewer pipe;

n corresponds to 5; and

said plurality of defects are deposits, cross-sectional reductions, misalignments, infiltration, and cracks.

12. A method as claimed in claim 11, wherein deposits, cross-sectional reductions, and milsalignments correspond to a first set of image analysis techniques, infiltration corresponds to a second set of image analysis techniques, and cracks correspond to a third set of image analysis techniques.

13. A method as claimed in claim 12, wherein said first set of image analysis techniques comprises the operations of inversion, dilation, background subtraction, thresholding, segmentation, and analysis.

14. A method as claimed in claim 12, wherein said second set of image analysis techniques comprises the operations of dilation, background subtraction, thresholding, segmentation, and analysis.

15. A method as claimed in claim 12, wherein said third set of image analysis techniques comprises the operations of background subtraction, edge detection, dilation, thresholding, and analysis.

16. A method as claimed in claim 15, wherein said videotape is digitized.

17. A method as claimed in claim 1, wherein said neural networks are back-propagation neural networks.

18. A method as claimed in claim 1, wherein said acquiring an image comprises using a closed circuit television camera and a videotape.

19. A method as claimed in claim 1, further comprising determining a position of said objects in said item under inspection.

20. A method for detecting a selected defect in an item under inspection comprising:

acquiring an image of said item;

providing a neural network for detecting said selected defect, wherein said selected defect is selected from at least one of deposits, cross-sectional reductions, misalignments, infiltration, and cracks;

selecting a set of image analysis techniques as a function of said selected defect;

processing said image according to said selected set of image analysis techniques for said selected defect to produce a processed image having objects isolated from an image background of said image;

inputting said processed image to said neural network to obtain information corresponding to said selected defect.

21. A method as claimed in claim 20, further comprising issuing a report based on outputs produced by said neural network.

22. A method as claimed in claim 21, further comprising recommending a rehabilitation technique based on said report and a set of attributes of said item under inspection.

23. A method as claimed in claim 22, wherein said attributes are part of a group comprising technical requirements, contractual requirements, and cost effectiveness.

24. A method as claimed in claim 22, wherein a plurality of rehabilitation techniques are recommended.

25. A method as claimed in claim 24, further comprising ranking said plurality of recommended rehabilitation techniques.

26. A method as claimed in claim 20, wherein said providing a neural network further comprises providing a set of neural networks, said set of neural networks being used for counter-checking results.

27. A method as claimed in claim 26, wherein said set of neural networks comprises three neural networks.

28. A method as claimed in claim 20, wherein deposits, cross-sectional reductions, and misalignments correspond to a first set of image analysis techniques, infiltration corresponds to a second set of image analysis techniques, and cracks correspond to a third set of image analysis techniques.

29. A method as claimed in claim 28, wherein said first set of image analysis techniques comprises the operations of inversion, dilation, background subtraction, thresholding, segmentation, and analysis.

30. A method as claimed in claim 28, wherein said second set of image analysis techniques comprises the operations of dilation, background subtraction, thresholding, segmentation, and analysis.

31. A method as claimed in claim 28, wherein said third set of image analysis techniques comprises the operations of background subtraction, edge detection, dilation, thresholding, and analysis.

32. A method as claimed in claim 20, wherein said neural network is a back-propagation neural network.

33. A method as claimed in claim 20, wherein said acquiring an image comprises using a closed circuit television camera and a videotape.

34. A method as claimed in claim 33, wherein said videotape is digitized.

35. A method as claimed in claim 20, further comprising determining a position of said objects in said item under inspection.