Automatic image thresholding using Otsu's method and entropy weighting scheme for surface defect detection

Defect detection is one of the most important tasks and a challenging problem for industrial quality control. Among the available visual inspection techniques, automatic thresholding is a commonly used approach for defect detection because of the simplicity in terms of its implementation and computing. In this paper, we propose an automatic thresholding technique, which is an improvement in Otsu's method, using an entropy weighting scheme. The proposed method enables the detection of extremely small defect regions compared to the product surface area. Experimental results confirm the efficiency of the proposed system over other techniques.