Surface defect inspection method and surface defect inspection apparatus

Patent metadata

Field	Value
Publication number	US-10859507-B2
Application number	US-201816495228-A
Country	US
Kind code	B2
Filing date	Feb 28, 2018
Priority date	Mar 21, 2017
Publication date	Dec 8, 2020
Grant date	Dec 8, 2020

Abstract

A surface defect inspection method includes: acquiring an original image by capturing an image of a subject of an inspection; generating texture feature images by applying a filtering process using spatial filters to the original image; generating a feature vector at each position of the original image, by extracting a value at a corresponding position from each of the texture feature images, for each of the positions of the original image; generating an abnormality level image representing an abnormality level for each position of the original image, by calculating, for each of the feature vectors, an abnormality level in a multi-dimensional distribution formed by the feature vectors; and detecting a part having the abnormality level that is higher than a predetermined level in the abnormality level image as a defect portion or a defect candidate portion.

First claim

The invention claimed is:

1. A surface defect inspection method comprising:
an image input step of acquiring an original image by capturing an image of a subject of an inspection;
a texture feature image generating step of generating texture feature images by applying a filtering process using spatial filters to the original image, wherein the texture feature image generating step includes a process of generating an additional texture feature image by applying the filtering process using the spatial filters to an image that is a reduction of the original image or to an image that is a reduction of one of the texture feature images;
a texture feature extracting step of generating a feature vector at each position of the original image, by extracting a value at a corresponding position from each of the texture feature images and the additional texture feature image, for each of the positions of the original image;
an abnormality level calculating step of generating an abnormality level image representing an abnormality level for each position of the original image, by calculating, for each of the feature vectors, an abnormality level in a multi-dimensional distribution formed by the feature vectors; and
a detecting step of detecting a part having the abnormality level that is higher than a predetermined level in the abnormality level image as a defect portion or a defect candidate portion.

2. The surface defect inspection method according to claim 1 , wherein a direction for reducing the original image or for reducing the one of the texture feature images includes a direction in parallel with a linear defect that is to be detected.

3. The surface defect inspection method according to claim 2 , wherein the spatial filters are achieved by wavelet conversion.

4. The surface defect inspection method according to claim 3 , wherein the spatial filters includes a Gabor filter.

5. The surface defect inspection method according to claim 4 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

6. The surface defect inspection method according to claim 3 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

7. The surface defect inspection method according to claim 2 , wherein the spatial filters includes a Gabor filter.

8. The surface defect inspection method according to claim 7 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

9. The surface defect inspection method according to claim 2 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

10. The surface defect inspection method according to claim 1 , wherein the spatial filters are achieved by wavelet conversion.

11. The surface defect inspection method according to claim 10 , wherein the spatial filters includes a Gabor filter.

12. The surface defect inspection method according to claim 11 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

13. The surface defect inspection method according to claim 10 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

14. The surface defect inspection method according to claim 1 , wherein the spatial filters includes a Gabor filter.

15. The surface defect inspection method according to claim 14 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

16. The surface defect inspection method according to claim 1 , wherein Mahalanobis distance is used as an abnormality level in the multi-dimensional distribution formed by the feature vectors.

17. A surface defect inspection apparatus comprising:
an image capturing unit configured to capture a subject of an inspection;
an image input unit configured to acquire an original image of the subject of the inspection, the original image being captured by the image capturing unit;
a texture feature image generating unit configured to generate texture feature images by applying a filtering process using spatial filters to the original image, wherein the texture feature image generating unit is configured to generate an additional texture feature image by applying the filtering process using the spatial filters to an image that is a reduction of the original image or to an image that is a reduction of one of the texture feature images;
a texture feature extracting unit configured to generate a feature vector at each position of the original image by extracting a value at a corresponding position from each of the texture feature images additional texture feature image, for each of the positions of the original image;
an abnormality level calculating unit configured to generate an abnormality level image representing an abnormality level for each position of the original image, by calculating, for each of the feature vectors, an abnormality level in a multi-dimensional distribution formed by the feature vectors; and
a detecting unit configured to detect a part having the abnormality level that is higher than a predetermined level in the abnormality level image as a defect portion or a defect candidate portion.

Assignees

• Jfe Steel Corp

Inventors

• Koshihara Takahiro
• Umegaki Yoshiyuki
• Oshige Takahiko

Classifications

• G01N21/8983 Primary

  for testing textile webs, i.e. woven material · CPC title

• G06V10/267

  by performing operations on regions, e.g. growing, shrinking or watersheds · CPC title

• G06T7/0004 Primary

  Industrial image inspection · CPC title

• G06V10/758

  Involving statistics of pixels or of feature values, e.g. histogram matching · CPC title

• G06V10/449

  Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters · CPC title