Application of multi-scale feature extraction to surface defect classification of hot-rolled steels

Abstract

Feature extraction is essential to the classification of surface defect images. The defects of hot-rolled steels distribute in different directions. Therefore, the methods of multi-scale geometric analysis (MGA) were employed to decompose the image into several directional subbands at several scales. Then, the statistical features of each subband were calculated to produce a high-dimensional feature vector, which was reduced to a lower-dimensional vector by graph embedding algorithms. Finally, support vector machine (SVM) was used for defect classification. The multi-scale feature extraction method was implemented via curvelet transform and kernel locality preserving projections (KLPP). Experiment results show that the proposed method is effective for classifying the surface defects of hot-rolled steels and the total classification rate is up to 97.33%.

Keywords

hot rolling / strip metal / surface defects / classification / feature extraction

References

[1] Song Q., Xu K., Xu J. Recognition of surface defects on medium and heavy plates based on structure spectrum. J. Univ. Sci. Technol. Beijing, 2007, 29(3): 342.

[2] K. Xu, K.P. Chen, and C.L. Yang, Application of frac tion dimensions based on the optimized scale to classification of surface defects on hot-rolled steel strips, Metall. Equip., 2008, No. 2, p. 1.

[3] Li J., Shi J., Chang T.S. On-line seam detection in rolling processes using snake projection and discrete wavelet transform. J. Manuf. Sci. Eng., 2007, 129(5): 926.

[4] Chuang W. L., Chen C.H., Yen J.Y., Hsu Y.L. Using MPCA of spectra model for fault detection in a hot strip mill. J. Mater. Process. Technol., 2009, 209(8): 4162.

[5] J. K. Romberg, M.B. Wakin, and R.G. Baraniuk, Multiscale geometric image processing, [in] Proceedings of the International Society for Optical Engineering, Lugano, 2003, p. 1265.

[6] Donoho D. L., Huo X.M. Beamlets and Multiscale Image Analysis, 2001, California, Standford University

[7] Donoho D. L., Duncan M.R. Digital Curvelet Transform: Strategy, Implementation and Experiments, 1999, California, Stanford University

[8] M. N. Do and M. Vetterli, Contourlets: a directional multiresolution image representation, [in] Proceedings of the International Conference on Image Processing, Rochester, 2002, p. I–357.

[9] Candès E. J. Ridgelet: Theory and Application, 1998, California, Stanford University, 13.

[10] Candès E. J. Monoscale Ridgelets for the Representation of Images with Edges, 1999, California, Stanford University, 1.

[11] M. J. Fadili and J.L. Starck, Curvelets and ridgelets [R/OL], [2011-12-19], http://jstarck.free.fr/curvencyclop09.pdf.

[12] Candès E. J., Demanet L., Donoho D.L., Ying L. Fast discrete curvelet transforms. Multiscale Model. Simul., 2006, 5(3): 861.

[13] S. C. Yan, D. Xu, B.Y. Zhang, and H.J. Zhang, Graph embedding: a general framework for dimensionality reduction, [in] Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Diego, 2005, p. 830.

[14] Tenenbaum J. B., de Silva V., Langford J.C. A global geometric framework for nonlinear dimensionality reduction. Science, 2000, 290, 2319.

[15] Roweis S. T., Saul L.K. Nonlinear dimensionality reduction by locally linear embedding. Science, 2000, 290, 2323.

[16] M. Belkin and P. Niyogi, Laplacian Eigenmaps and spectral techniques for embedding and clustering, [in] Advances in Neural Information Processing Systems, Cambridge, 2002, p. 585.

[17] X. F. He and P. Niyogi, Locality preserving projections, [in] Advances in Neural Information Processing Systems, Vancouver, 2003.