An improved Otsu algorithm is developed for the image segmentation of the printed circuit board(PCB).The fuzzy area is separated first by using double thresholds based on Otsu theory.Taking into account the characteristic of PCB images and making use of the ability of the fuzzy set theory depicting ambiguity
we construct the membership function of the images.Finally the algorithm based on fuzzy set and Otsu theory is formed.Many experiments were conducted on different images
and the results show that the algorithm has good segmentation performance for PCB images.
Ng H H.Automatic thresholding for defect detection[J].Pattern Recognition Letters,2006,27:1644-1649.
付忠良.图像阈值选取方法-Otsu方法的推广[J].计算机应用,2000,20(5):37-39.Fu Zhongliang.The selecting method of image threshold-the popularization of otsu[J].Application Research ofComputers,2000,20(5):37-39.(in Chinese)
王新洲,史文中,王树良.模糊空间信息处理[M].武汉:武汉大学出版社,2003.
Li Zhenquan.Suitability of fuzzy reasoning methods[J].Fuzzy Sets and System,1998,108(3):299-311.