Visual Inspection Method of Ceramic Bottle Surface Defects Based on Niblack Optimization

Abstract

In order to improve the efficiency and accuracy of the defect detection of ceramic bottle production line, a method of visual inspection of ceramic bottle surface defects based on threshold segmentation Niblack optimization algorithm is been proposed. The CLAHE interpolation image enhancement algorithm is used to improve the image contrast, and then the local threshold is obtained by calculating the local mean and standard deviation in the domain window to realize the segmentation of the defective region. Then the minimum gray value difference is introduced and whether there are defects in images are determined to the center pixel gray value range. The characteristics of the defect target are extracted, and the defect classifier is used to classify the defects according to the principle of support vector machine. The experimental results show that the optimization algorithm can realize the effective segmentation of the defect area and reduce the false detection rate by reducing the influence of background noise on the segmentation results with an accuracy rate of 90.8%.