Abstract

In this paper we deal with the problem of detecting and segmenting objects in textured dark-field digital imagery for automated visual-inspection applications. We first present a technique for correcting optical shading effects in conventional dark-field microscopy. After compensating for possible imperfections in the optical setting we address the problem of segmenting objects (defects) in textured dark-field images. The technique that we will follow is based on a sequential application of local operators, which serves the purpose of clustering the object and the background gray levels. This procedure can be considered an extension of average-thresholding-type techniques. Both algorithms for shading correction and object segmentation have fast implementations in general-purpose image-processing pipeline architectures, and therefore they are appealing to real-time computer vision applications. Computational examples showing the appropriateness of the shading-correction procedure as well as the effectiveness of the segmentation wil be discussed.

© 1985 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
    [CrossRef]
  2. L. C. Martin, The Theory of the Microscope (American Elsevier, New York, 1966).
  3. In this paper all histograms have been smoothed by applying a local averaging operator every three entries.
  4. I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).
  5. K. S. Fu, J. K. Mu, “A survey on image segmentation,” Pattern Recog. 13, 3–16 (1981).
    [CrossRef]
  6. J. S. Wezska, “A survey of threshold selection techniques,”Comput. Graphics Image Process. 7, 259–265 (1978).
    [CrossRef]
  7. B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
    [CrossRef]
  8. D. B. Cooper, F. Sung, “Multiple-window parallel adaptive boundary finding in computer vision,” IEEE Trans. Pattern Anal. Mach. Intell. PAMI-5, 299–265 (1983).
    [CrossRef]
  9. S. Horowitz, T. Pavlidis, “Picture segmentation by a tree traversal algorithm,” J. Assoc. Comput. Mach. 23, 368–396 (1976).
    [CrossRef]
  10. R. Nevatia, Machine Perception (Prentice-Hall, Englewood Cliffs, N.J., 1982).
  11. R. Nevatia, “Locating object boundaries in textured environments,” IEEE Trans. Comput. C-25, 1170–1175 (1976).
    [CrossRef]
  12. J. L. C. Sanz, M. D. Flickner, “Computing minima and maxima of digital images in pipeline image processing systems without hardware comparators,” Proc. IEEE (to be published).

1983 (1)

D. B. Cooper, F. Sung, “Multiple-window parallel adaptive boundary finding in computer vision,” IEEE Trans. Pattern Anal. Mach. Intell. PAMI-5, 299–265 (1983).
[CrossRef]

1981 (1)

K. S. Fu, J. K. Mu, “A survey on image segmentation,” Pattern Recog. 13, 3–16 (1981).
[CrossRef]

1979 (1)

B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
[CrossRef]

1978 (1)

J. S. Wezska, “A survey of threshold selection techniques,”Comput. Graphics Image Process. 7, 259–265 (1978).
[CrossRef]

1976 (2)

S. Horowitz, T. Pavlidis, “Picture segmentation by a tree traversal algorithm,” J. Assoc. Comput. Mach. 23, 368–396 (1976).
[CrossRef]

R. Nevatia, “Locating object boundaries in textured environments,” IEEE Trans. Comput. C-25, 1170–1175 (1976).
[CrossRef]

1975 (1)

L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
[CrossRef]

Cooper, D. B.

D. B. Cooper, F. Sung, “Multiple-window parallel adaptive boundary finding in computer vision,” IEEE Trans. Pattern Anal. Mach. Intell. PAMI-5, 299–265 (1983).
[CrossRef]

Davis, L. S.

B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
[CrossRef]

L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
[CrossRef]

Dinstein, I.

I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).

Flickner, M. D.

J. L. C. Sanz, M. D. Flickner, “Computing minima and maxima of digital images in pipeline image processing systems without hardware comparators,” Proc. IEEE (to be published).

Fu, K. S.

K. S. Fu, J. K. Mu, “A survey on image segmentation,” Pattern Recog. 13, 3–16 (1981).
[CrossRef]

Horowitz, S.

S. Horowitz, T. Pavlidis, “Picture segmentation by a tree traversal algorithm,” J. Assoc. Comput. Mach. 23, 368–396 (1976).
[CrossRef]

Lam, T. D.

I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).

Martin, L. C.

L. C. Martin, The Theory of the Microscope (American Elsevier, New York, 1966).

Merkle, F.

I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).

Mu, J. K.

K. S. Fu, J. K. Mu, “A survey on image segmentation,” Pattern Recog. 13, 3–16 (1981).
[CrossRef]

Nevatia, R.

R. Nevatia, “Locating object boundaries in textured environments,” IEEE Trans. Comput. C-25, 1170–1175 (1976).
[CrossRef]

R. Nevatia, Machine Perception (Prentice-Hall, Englewood Cliffs, N.J., 1982).

Pavlidis, T.

S. Horowitz, T. Pavlidis, “Picture segmentation by a tree traversal algorithm,” J. Assoc. Comput. Mach. 23, 368–396 (1976).
[CrossRef]

Rosenfeld, A.

B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
[CrossRef]

L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
[CrossRef]

Sanz, J. L. C.

J. L. C. Sanz, M. D. Flickner, “Computing minima and maxima of digital images in pipeline image processing systems without hardware comparators,” Proc. IEEE (to be published).

Schachter, B. J.

B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
[CrossRef]

Sung, F.

D. B. Cooper, F. Sung, “Multiple-window parallel adaptive boundary finding in computer vision,” IEEE Trans. Pattern Anal. Mach. Intell. PAMI-5, 299–265 (1983).
[CrossRef]

Weszka, J. S.

L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
[CrossRef]

Wezska, J. S.

J. S. Wezska, “A survey of threshold selection techniques,”Comput. Graphics Image Process. 7, 259–265 (1978).
[CrossRef]

Wong, K. Y.

I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).

Comput. Graphics Image Process. (1)

J. S. Wezska, “A survey of threshold selection techniques,”Comput. Graphics Image Process. 7, 259–265 (1978).
[CrossRef]

IEEE Trans. Comput. (1)

R. Nevatia, “Locating object boundaries in textured environments,” IEEE Trans. Comput. C-25, 1170–1175 (1976).
[CrossRef]

IEEE Trans. Pattern Anal. Mach. Intell. (1)

D. B. Cooper, F. Sung, “Multiple-window parallel adaptive boundary finding in computer vision,” IEEE Trans. Pattern Anal. Mach. Intell. PAMI-5, 299–265 (1983).
[CrossRef]

IEEE Trans. Systems Man Cybern. (1)

L. S. Davis, A. Rosenfeld, J. S. Weszka, “Region extraction by averaging and thresholding,” IEEE Trans. Systems Man Cybern. SMC-5, 383–388 (1975).
[CrossRef]

J. Assoc. Comput. Mach. (1)

S. Horowitz, T. Pavlidis, “Picture segmentation by a tree traversal algorithm,” J. Assoc. Comput. Mach. 23, 368–396 (1976).
[CrossRef]

Pattern Recog. (2)

B. J. Schachter, L. S. Davis, A. Rosenfeld, “Some experiments in image segmentation by clustering of local feature values,” Pattern Recog. 11, 19–28 (1979).
[CrossRef]

K. S. Fu, J. K. Mu, “A survey on image segmentation,” Pattern Recog. 13, 3–16 (1981).
[CrossRef]

Other (5)

J. L. C. Sanz, M. D. Flickner, “Computing minima and maxima of digital images in pipeline image processing systems without hardware comparators,” Proc. IEEE (to be published).

R. Nevatia, Machine Perception (Prentice-Hall, Englewood Cliffs, N.J., 1982).

L. C. Martin, The Theory of the Microscope (American Elsevier, New York, 1966).

In this paper all histograms have been smoothed by applying a local averaging operator every three entries.

I. Dinstein, F. Merkle, T. D. Lam, K. Y. Wong, “Imaging system response linearization and shading correction,” IBM Res. Rep. RJ4044 (45302);Opt. Eng. (to be published).

Cited By

OSA participates in CrossRef's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (8)

Fig. 1
Fig. 1

Dark-field imaging.

Fig. 2
Fig. 2

(a) Dark-field digital image without shading correction. (b) Histogram of image (a). (c) Four different thresholds applied to image (a), showing the clear effect of shading.

Fig. 3
Fig. 3

(a) Optomechanically averaged image. (b) Histogram of an optomechanically averaged image showing the shading phenomenon.

Fig. 4
Fig. 4

Histogram of an optomechanically averaged image after shading correction.

Fig. 5
Fig. 5

(a) Dark-field digital image after shading correction. (b) Histogram of image (a). (c) Four different thresholds applied to image (a) showing the compensation for the shading problem.

Fig. 6
Fig. 6

(a) Dark-field digital image averaged using a 5 × 5 window. (b) Histogram of image (a). (c) Thresholded image (a). (d) Image processed with new algorithm after three iterations. (e) Histogram of image (d). A good resolution between clusters is obtained. (f) Thresholded image (d).

Fig. 7
Fig. 7

(a) Dark-field digital image averaged using a 5 × 5 window. (b) Histogram of image (a). No threshold selection is possible. (c)–(f) Image (a) processed with new algorithm: first two iterations and their corresponding histograms. (g)–(j) Image (a) processed with new algorithm: third and fourth iterations and their corresponding histograms. (k) Thresholded image (i).

Fig. 8
Fig. 8

(a) Digital dark-field image after averaging. (b) Histogram of image (a). A poor resolution between clusters is observed. (c) Thresholded image (a). (d) Image (a) processed with new algorithm after five iterations. (e) Histogram of image (d). An improved resolution between clusters is obtained. (f) Thresholded image (d).

Equations (4)

Equations on this page are rendered with MathJax. Learn more.

CORIMG ( i , j ) = IMG ( i , j ) CF ( i , j )
IMG ( i , j ) = ORGIMG ( i , j ) BIAS ( i , j ) for all pixels ( i , j ) .
CF ( i , j ) = MAX / REFIMG ( i , j ) for all pixels ( i , j ) ,
min ( a , b ) = ( | a b | + a + b ) 2 ,

Metrics