Defect analysis system

• US 4,484,081 A
• Filed: 03/22/1984
• Issued: 11/20/1984
• Est. Priority Date: 09/19/1980
• Status: Expired due to Fees

• Current Assignee: TRW Inc. (Northrop Grumman Corporation)
• Original Assignee: TRW Inc. (Northrop Grumman Corporation)
• Inventors: Barrett, John S., Brabston, Donald C. Jr., Nelson, Steven P., Camana, Peter C., Cornyn, William S. Jr., McAferty, James M., Schy, Stuart T.
• Primary Examiner(s): Nelms, David C.
• Application Number:
  US06/592,734
• Time in Patent Office: 243 Days
• Field of Search: 250/562, 250/563, 250/572, 250/578, 358/101, 358/106, 358/107, 356/237, 356/238, 356/430, 356/394, 364/507, 364/506, 364/552, 364/554, 364/576, 364/577, 364/551
• US Class Current: 250/559.46
• CPC Class Codes: G01N 2021/646 Detecting fluorescent inhomogeneities at a position, e.g. for detecting defects
• CPC Class Codes: G01N 21/89 in moving material, e.g. running paper or textiles G01N21/90, G01N21/91, G01N21/94 take precedence
• CPC Class Codes: G01N 21/9515 Objects of complex shape, e.g. examined with use of a surface follower device measuring contours and curvatures G01B11/24
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/30164 Workpiece; Machine component
• CPC Class Codes: G06T 7/0006 using a design-rule based approach

Abstract

Disclosed is an apparatus, and a related method, for the automatic detection of defects in manufactured parts, such as turbine blades and the like. The apparatus includes a video camera, a thresholding module to provide a binary image of a frame of video data, a region growing module to define anomalous regions by associating contiguous picture elements indicative of surface anomalies on the part, and defect analysis modules for determining if a rejectable defect exists, based on a comparison between statistics with respect to the anomalous regions and predetermined rejection criteria. Rejection may be based on region size, region shape, or region position and orientation with respect to neighboring regions. Thus, a region may be rejected if it is nearly collinear with another region, even though neither region may itself be characterized as a rejectable region.

Claims

1. A defect analysis system for examining a part to determine whether defects are present, said system comprising:

first means for obtaining electrical signals representative of the condition of a plurality of discrete picture elements in a field of view of the part presented for analysis;

thresholding means coupled to said first means for generating a set of electrical signals indicative of picture elements of an anomalous type distinct from the other picture elements;

region growing means for associating contiguous picture elements of the anomalous type to define multiple regions of anomalous picture elements in a single raster scan of the field of view being analyzed;

region statistics generation means for generating statistics relative to the size, shape and orientation of the anomalous regions during the same single raster scan; and

region analysis means for comparing the region statistics with predetermined rejection parameters to determine whether a rejectable defect exists in the part.

2. A defect analysis system as set forth in claim 1, wherein:

said region statistics generation means includes means for recording the position, area, length-to-width ratio and orientation of each region.

3. A defect analysis system as set forth in claim 2, wherein said region analysis means includes:

means for comparing the area of a region with a first preselected limit wherein regions smaller in area than the first preselected limit are disregarded; and

means for comparing the area of a region with a second preselected minimum area, wherein regions larger than the second preselected minimum area are designated as rejectable defects.

4. A defect analysis system as set forth in claim 2, wherein said region analysis means includes:

means for comparing the length-to-width ratio of a region with a preselected threshold ratio, wherein regions having a length-to-width ratio higher than the threshold ratio are designated as rejectable defects.

5. A defect analysis system as set forth in claim 2, wherein said region analysis means includes:

means for comparing the spacing between and orientation of two regions with respective preselected spacing and orientation parameters to determine whether two regions together constitute a rejectable defect.

6. A defect analysis system as set forth in claim 5, wherein said means for comparing include:

means for determining the centers of mass of the two regions;

means for comparing the distance between the centers of mass of the two regions, with a preselected multiple of the length of one of the regions;

means for comparing the difference in angular orientations of the regions with a preselected angle; and

means for comparing the difference between the angular orientation of one region and the angular orientation of a line joining the centers of mass of both regions, with a preselected angle, wherein a rejectable defect occurs when the necessary distance and angle values are smaller than the corresponding preselected values.

7. A defect analysis system as set forth in claim 1, wherein:

said system further includes a source of radiation directed to the part being examined;

the condition of the picture elements that renders anomalous elements distinct is the intensity of radiation reflected from each element; and

said thresholding means includes

histogram calculation means for accumulating counts of picture elements for corresponding levels of radiation intensity,

means for determining from the accumulated counts a threshold level of intensity above which a picture element will be considered to be anomalous, and

image transformation means for comparing each picture element with the threshold level and thereby generating a binary image of the field of view.

8. A deflect analysis system as set forth in claim 7, wherein said histogram calculation means and said image transformation means are implemented in hard-wired form to increase the speed of operation.

9. A surface defect analysis system for examining a part to determine whether structural defects are present, said system comprising:

a source of light directed to the part being examined;

video camera means for obtaining analog electrical signals representative of the intensity of light emitted by discrete picture elements in an image frame of the part presented for analysis;

thresholding means coupled to said video camera means, for generating binary electrical signals indicative of anomalous picture elements having light intensities substantially different from the other picture elements;

region growing means for associating contiguous anomalous picture elements to define multiple regions of anomalous picture elements in a single raster scan of the image frame presented for analysis;

region statistics generation means for generating statistics relative to the size, shape and orientation of the anomalous regions during the same single raster scan; and

region analysis means for comparing the region statistics with predetermined rejection parameters, to determine whether a rejectable defect is present in the part.

10. A surface defect analysis system as set forth in claim 9, wherein:

said region statistics generation means includes means for recording the position, area, length-to-width ratio and orientation of each region.

11. A surface defect analysis system as set forth in claim 10, wherein said region analysis means includes:

means for comparing the area of a region with a first preselected limit, wherein regions smaller in area than the first preselected limit are disregarded; and

means for comparing the area of a region with a second preselected limit wherein regions larger in area than the second preselected limit are designated as rejectable defects.

12. A surface defect analysis system as set forth in claim 10, wherein said region analysis means includes:

means for comparing the length-to-width ratio of a region with a preselected threshold ratio, wherein regions having a length-to-width ratio higher than the threshold ratio are designated as rejectable defects.

13. A surface defect analysis system as set forth in claim 10, wherein said region analysis means includes:

means for comparing the spacing between and orientation of two regions with respective preselected spacing and orientation parameters to determine whether two regions together constitute a rejectable defect.

14. A surface defect analysis system as set forth in claim 13, wherein said means for comparing include:

means for comparing the distance between the centers of mass of the two regions, with a preselected multiple of the length of one of the regions;

means for comparing the difference in angular orientations of the regions with a preselected angle;

means for comparing the difference between the angular orientation of one region and the angular orientation of a line joining the centers of mass of both regions, within a preselected angle, wherein a rejectable defect occurs when the necessary distance and angle values are smaller than the corresponding preselected values.

15. A surface defect analysis system as set forth in claim 9, wherein said thresholding means includes

histogram calculation means for accumulating counts of picture elements for corresponding levels of light intensity,

means for determining from the accumulated counts a threshold level of intensity above which a picture element will be considered to be anomalous, and

image transformation means for comparing each picture element with the threshold level and thereby generating a binary frame image.

16. A surface defect analysis system as set forth in claim 15, wherein said histogram calculation means and said image transformation means are implemented in hard-wired form to increase the speed of operation.

17. A surface defect analysis system as set forth in claim 15, wherein said means for determining a threshold level includes:

means for comparing the ratio of two adjacent histogram values with a preselected ratio; and

means for designating a histogram value the threshold level based on the results obtained from two successive comparison operations by said means for comparing.

18. A surface defect analysis system as set forth in claim 17, wherein said preselected ratio is 0.5.

19. A method for detection of defects in a manufactured part, comprising the steps of:

obtaining electrical signals representative of a plurality of picture elements in a field of view of a part presented for inspection;

transforming these signals into a set of binary electrical signals indicative of anomalous picture elements that are optically distinct from other picture elements;

associating contiguous anomalous picture elements to define multiple regions of anomalous picture elements in a single raster scan of the field of view presented for inspection;

generating statistics relative to physical characteristics of each region during the same single raster scan; and

comparing region statistics with predetermined rejection parameters to determine whether a rejectable defect is present.

20. A method as set forth in claim 19, wherein said step of transforming the signals includes:

generating a histogram of the entire image frame to define the relationship between light intensity and numbers of picture elements;

calculating an optimum threshold level from the histogram; and

generating a binary frame image by comparing each picture element intensity with the threshold level.

21. A method as set forth in claim 19, wherein:

said step of generating statistics includes recording the size, position, length-to-width ratio, and angular orientation of each region.

22. A method as set forth in claim 21, wherein:

said step of comparing region statistics with predetermined rejection parameters includes, comparing the area of each region with a preselected limit above which the region will be designated a defect.

23. A method as set forth in claim 21, wherein:

said step of comparing region statistics with predetermined rejection parameters includes comparing the length-to-width ratio of each region with a preselected value, above which the region will be designated a defect.

24. A method as set forth in claim 21, wherein said step of comparing region statistics with predetermined rejection parameters includes:

comparing the distance to each neighboring region with a preselected relative distance; and

comparing the angular orientation of each region with that of each neighboring region, whereby a rejectable defect will be found if two regions are relatively close together and are substantially collinear.