Inspection of Surface Defects in Copper Strip Based on Machine Vision

Though copper products are important raw materials in industrial production, there is little domestic research focused on copper strip surface defects inspection based on automated visual inspection. According to the defect image characteristics on copper strips surface, a defect detection algorithm is proposed on the basis of wavelet-based multivariate statistical approach. First, the image is divided into several sub-images, and then each sub-image is further decomposed into multiple wavelet processing units. Then each wavelet processing unit is decomposed by 1-D db4 wavelet function. Then multivariate statistics of Hotelling T-2 are applied to detect the defects and SVM is used as defect classifier. Finally, the defect detection performance of the proposed approach is compared with traditional method based on grayscale. Experimental results show that the proposed method has better performance on identification, especially its application in the ripple defects can achieve 96.7% accuracy, which was poor in common algorithms.