Binary Gabor pattern (BGP) descriptor and principal component analysis (PCA) for steel surface defects classification

Efficient surface defect classification is one of the most important factors to achieve online quality inspection for hot-rolled strip steels. It is extremely challenging owing to its localization on a large surface, various defect appearance, large scale changes of defects, and random distribution. Therefore, in this paper, we proposed an efficient system for steel surface defects classification that can attain excellent classification accuracy. The presented system extracts local texture features from defect images, by application of the binary Gabor pattern (BGP) descriptor used for the first time on the steel surface defects classification. Then, a dimensionality reduction procedure, based on the principal component analysis (PCA) is employed to obtain compact representation of the defects image. Lastly, SVM multiclass classifier is utilized to give the final decision. A set of experiments was conducted on the NEU Surface Defects database to investigate the performance of the proposed system. The results obtained demonstrate the effectiveness of the proposed approach for steel surface defects classification.