1 School of Electronic and Information Engineering, University of Science and Technology Liaoning, Anshan 114051, Liaoning, China 2 Department of Electrical Engineering, Lakehead University, Thunder Bay P7B 5E1, ON, Canada 3 State Key Laboratory of Robotics and System HIT, Harbin 150001, Heilongjiang, China
Multi-class classification method for strip steel surface defects based on support vector machine with adjustable hyper-sphere
1 School of Electronic and Information Engineering, University of Science and Technology Liaoning, Anshan 114051, Liaoning, China 2 Department of Electrical Engineering, Lakehead University, Thunder Bay P7B 5E1, ON, Canada 3 State Key Laboratory of Robotics and System HIT, Harbin 150001, Heilongjiang, China
摘要 Focusing on strip steel surface defects classification, a novel support vector machine with adjustable hyper-sphere (AHSVM) is formulated. Meanwhile, a new multi-class classification method is proposed. Originated from support vector data description, AHSVM adopts hyper-sphere to solve classification problem. AHSVM can obey two principles: the margin maximization and inner-class dispersion minimization. Moreover, the hyper-sphere of AHSVM is adjustable, which makes the final classification hyper-sphere optimal for training dataset. On the other hand, AHSVM is combined with binary tree to solve multi-class classification for steel surface defects. A scheme of samples pruning in mapped feature space is provided, which can reduce the number of training samples under the premise of classification accuracy, resulting in the improvements of classification speed. Finally, some testing experiments are done for eight types of strip steel surface defects. Experimental results show that multi-class AHSVM classifier exhibits satisfactory results in classification accuracy and efficiency.
Abstract：Focusing on strip steel surface defects classification, a novel support vector machine with adjustable hyper-sphere (AHSVM) is formulated. Meanwhile, a new multi-class classification method is proposed. Originated from support vector data description, AHSVM adopts hyper-sphere to solve classification problem. AHSVM can obey two principles: the margin maximization and inner-class dispersion minimization. Moreover, the hyper-sphere of AHSVM is adjustable, which makes the final classification hyper-sphere optimal for training dataset. On the other hand, AHSVM is combined with binary tree to solve multi-class classification for steel surface defects. A scheme of samples pruning in mapped feature space is provided, which can reduce the number of training samples under the premise of classification accuracy, resulting in the improvements of classification speed. Finally, some testing experiments are done for eight types of strip steel surface defects. Experimental results show that multi-class AHSVM classifier exhibits satisfactory results in classification accuracy and efficiency.
CHU Mao-Xiang,XIAO-PING -LIU,GONG Rong-Fen, et al. Multi-class classification method for strip steel surface defects based on support vector machine with adjustable hyper-sphere[J]. Journal of Iron and Steel Research International, 2018, 25(7): 706-716.
[1]
Y.J. Liu, J.Y. Kong, X.D. Wang, F.Z. Jiang, in: 3rd International Conference on Advanced Computer Theory and Engineering, IEEE, Piscataway, NJ, 2010, pp. 189-192.
[2]
F. Pernkopf, P. O'Leary, NDT & E International 36 (2003) No. 8, 609-617.
[3]
S.Q. Guan, ISIJ International 55 (2015) No. 9, 1950-1955.
[4]
K.C. Song, S.P. Hu, Y.H. Yan, J. Li, ISIJ International 54 (2014) No. 11, 2598-2607.
[5]
Y. Zhang, W.W. Liu, Z.T. Xing, Y.H. Yan, Journal of Northeastern University (Natural Science) 33 (2012) No. 2, 267-270.
2018, Vol. 25 
