System and method for automated fluorescent penetrant inspection

Patent metadata

Field	Value
Publication number	US-10054552-B1
Application number	US-201715716911-A
Country	US
Kind code	B1
Filing date	Sep 27, 2017
Priority date	Sep 27, 2017
Publication date	Aug 21, 2018
Grant date	Aug 21, 2018

Abstract

An example method of inspecting a part includes applying a penetrant dye to the part, the penetrant dye exhibiting a fluorescent color when subjected to light from a lighting device. A portion of the part is illuminated with light from the lighting device. An image of the portion of the part is automatically recorded with a camera while the portion is illuminated. An uncertainty metric for the image is automatically determined that is indicative of a likelihood that pixels in the image having the fluorescent color represent damage to the part. At least one of the part, lighting device, and camera are automatically adjusted based on the uncertainty metric being within a predefined range. The automatic recording, determining, and adjusting steps are iteratively repeated until the uncertainty metric is greater than the predefined range, or a predefined number of iterations have been performed for the portion of the part.

First claim

What is claimed is:

1. A method of inspecting a part, comprising:
applying a penetrant dye to a part, the penetrant dye exhibiting a fluorescent color when subjected to light from a lighting device;
illuminating a portion of the part with light from the lighting device;
automatically recording an image of the portion of the part with a camera while the portion is illuminated;
automatically determining an uncertainty metric for the image indicative of a likelihood that pixels in the image having the fluorescent color represent damage to the part;
automatically adjusting at least one of the part, the lighting device, and the camera based on the uncertainty metric being within a predefined range; and
iteratively repeating the automatic recording, determining, and adjusting steps until the uncertainty metric is greater than the predefined range, or a predefined number of iterations have been performed for the portion of the part.

2. The method of claim 1 , wherein the light is ultraviolet (UV) light.

3. The method of claim 1 , wherein said automatically adjusting at least one of the part, the lighting device, and the camera comprises automatically adjusting at least one of the following:
a relative position of the part with respect to one or both of the lighting device and the camera; and
an orientation of one or more of the part, the lighting device, and the camera.

4. The method of claim 1 , wherein said automatically adjusting at least one of the part, the lighting device, and the camera based comprises at least one of:
automatically adjusting one or more of a light intensity, light illumination area, and polarization of the lighting device; and
automatically adjusting an exposure setting of the camera to accommodate the adjustment to the lighting device.

5. The method of claim 1 , comprising:
determining that the portion of the part is damaged based on the uncertainty metric being greater than the predefined range; and
determining that the portion of the part is not damaged based on the uncertainty metric not exceeding the predefined range for any of the iterations.

6. The method of claim 5 , comprising:
based on a determination that the portion of the part is damaged, automatically depositing a damage marker on the part in proximity to the damage.

7. The method of claim 1 , wherein said determining an uncertainty metric for the image comprises:
converting the image, which uses a first color representation, to a second image that uses a different, second color representation;
performing a color detection to determine areas in the second image having the fluorescent color;
performing a pattern detection based on the color detection; and
determining a likelihood that a detected pattern depicts damage to the part based on pattern matching with a defect model.

8. The method of claim 7 , comprising:
situating a fiducial marker that exhibits the fluorescent color in close enough proximity to the portion of the part that it is recorded as part of the image;
wherein said performing a color detection is performed based on the fiducial marker in the image.

9. The method of claim 7 , wherein said performing a pattern detection comprises:
creating a skeleton image based on the color detection; and
performing a Hough transform of the skeletal image.

10. The method of claim 1 , comprising
situating a light filter between the part and a portion of either the camera or the lighting device when recording images, the light filter preventing one or more of certain non-UV light and light polarizations from passing through the filter.

11. The method of claim 1 , comprising
situating the camera, which is part of a borescope, inside the part when recording images of the portion of the part.

12. A part inspection system comprising:
a part onto which a penetrant dye is applied, the penetrant dye exhibiting a fluorescent color when subjected to light within a given frequency band;
a lighting device operable to project light within the given frequency band onto the part;
a camera; and
a controller configured to:
automatically record an image of a portion of the part with the camera while the portion is illuminated by the light within the given frequency band;
automatically determine an uncertainty metric for the image indicative of a likelihood that pixels in the image having the fluorescent color represent damage to the part;
automatically adjust at least one of the part, the lighting device, and the camera based on the uncertainty metric being within a predefined range; and
iteratively repeat the automatic recording, determining, and adjusting until the uncertainty metric is greater than the predefined range, or a predefined number of iterations have been performed for the portion of the part.

13. The part inspection system of claim 12 , comprising at least one robotic arm operable to move with six degrees of freedom, wherein said automatically adjusting at least one of the part, the light, and the camera based comprises manipulating the at least one robotic arm to adjust at least one of the following:
a relative position of the part with respect to one or both of the lighting device and the camera; and
an orientation of one or more of the part, the lighting device, and the camera.

14. The part inspection system of claim 12 , wherein to automatically adjust at least one of the part, the lighting device, and the camera, the controller is configured to perform at least one of:
automatically adjust one or more of a light intensity, light illumination area, and polarization of the lighting device; and
automatically adjust an exposure setting of the camera to accommodate the adjustment to the lighting device.

15. The part inspection system of claim 12 , wherein the controller is configured to:
determine that the portion of the part is damaged based on the uncertainty metric being greater than the predefined range; and
determine that the portion of the part is not damaged based on the uncertainty metric not exceeding the predefined range for any of the iterations.

16. The part inspection system of claim 15 , comprising a marking device, wherein the controller is configured to automatically deposit a damage marker on the part in proximity to the damage based on a determination that the portion of the part is damaged.

17. The part inspection system of claim 12 , wherein to determine an uncertainty metric for the image, the controller is configured to:
convert the image, which uses a first color representation, to a second image that uses a different, second color representation;
perform a color detection to determine areas in the second image having the fluorescent color;
perform a pattern detection based on the color detection; and
determine a likelihood that a detected pattern depicts damage to the part based on pattern matching with a defect model.

18. The part inspection system of claim 17 , wherein to perform a pattern detection, the controller is configured to:
create a skeleton image based on the color detection; and
perform a Hough transform of the skeletal image.

19. The part inspection system of claim 12 , wherein the given frequency band includes ultraviolet light, and the part inspection system comprises a light filter situated between the part and a portion of either the camera or the lighting device when the camera records an image of the part, the light filter operable to prevent one or more of certain non-UV light and light polarizations from passing through the filter.

Assignees

• United Technologies Corp

Inventors

• Frutuoso Sergio S
• Finn Alan Matthew
• Donskoy Gene B

Classifications

• G01N21/91 Primary

  using penetration of dyes, e.g. fluorescent ink · CPC title

• G01N2021/8887

  based on image processing techniques · CPC title

• G01N21/8851

  Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges (G01N21/8806 and G01N21/93 - G01N21/95692 take precedence; optical measurement of dimensions G01B11/00; optical scanning G02B26/10; image transformation G06T3/00; computerised image enhancement G06T5/00; image processing per se for flaw detection G06T7/0002) · CPC title

• G01N35/00584

  Control arrangements for automatic analysers · CPC title

• G01N21/8806

  Specially adapted optical and illumination features · CPC title