Image segmentation methods in problems of surface defect detection
A.P. Tsapaev, O.V. Kretinin

Full text of article: Russian language.

Abstract:
The analysis of image segmentation methods concerning to problem of surface defect detection is produced. Watershed method, normalized cut method and method of form spectrum change valuation (FSCV) are considered. Analysis is conducted on the model images and images of inspection object (images of inner tube surface). The conclusion about possibility of FSCV method application in visual surface inspection systems is drew.

Key words:
image processing, segmentation, visual inspection, defect, automation, pipes.

References:

  1. Ulyanov, A.N. The method and means of optoelectronic control the surface quality of sheet metal: a thesis for the degree of candidate of technical sciences. – Cherepovets. 2005. – 192p. – (In Russian)
  2. Starostin, D.A. The mathematical and simulation models of images of the surface of steel strip based Gibbs random fields: a thesis for the degree of candidate of technical sciences. – Cherepovets. 2003. – 158 p. – (In Russian)
  3. Image segmentation by clusters method and casual spasmodic structure algorithm: the comparative analysis / B.M. Mironov, A.N. Malov // Computer optics. – 2010. – V.34, N.1. – P. 132-137. – (In Russian).
  4. Vincent, P. Soille, Watersheds in Digital Space: An Efficient Algorithms based on Immersion Simulation // [J]-IEEE Transactions on Pattern Analysis and Machine Intelligence, – 1991. - 13, No.6, - P.583-598,.
  5.  Forsyth, David A., Computer vision. A modern approach.: / David A. Forsyth, Jean Ponce. Translated from English. – M.: Publishing House. «Williams», 2004. – (In Russian)
  6. Jitendra Malik, Serge Belongie, Txomas Leung, Jianbo Shi, «Contour and Texture Analysis for Image Segmentation», International Journal of Computer Vision, V.43, No.1, pp. 7-27, 2001.
  7. MATLAB Normalized Cuts Segmentation Code.        
    URL: http://www.cis.upenn.edu/~jshi/software/

    © 2009, IPSI RAS
    Institution of Russian Academy of Sciences, Image Processing Systems Institute of RAS, Russia, 443001, Samara, Molodogvardeyskaya Street 151; E-mail: ko@smr.ru; Phones: +7 (846 2) 332-56-22, Fax: +7 (846 2) 332-56-20