Strip surface defect detection algorithm based on background difference

Abstract

Detection of strip surface defects target is the focus and difficulty of Strip surface quality inspection system. Aimed at the problem resulting from uneven illumination and slowly-changing global illumination in the strip test site, the background difference algorithm is adopted in strip surface defect detection. Firstly normal strip surface image as the initial background is fast reconstructed through Gaussian-background reconstruction algorithm from the collected initial video sequences of strip surface, because of the grayscale differences between defect area and normal strip surface, the background is subtracted from the current frame image, and accurate extraction of the defect region is achieved after thresholding, labeling and defect-location processing. While taking advantage of the current frame to date information on background image can be well adapted to the slow changes in ambient illumination. Simulation result shows that the algorithm is not only simple in principle, but also performs well in robustness, rapidity and accuracy, thus it has laid the foundation for further recognition and classification.