Abstract: The iterative triclass thresholding technique is a parameter free technique. It is based on Otsuís method where the Otsuís threshold is obtained, which minimizes the intra-class variances of the segmented images, but instead of classifying the histogram of that image into two classes, the iterative method separates it into three classes. This triclass is defined as: 1) foreground class, 2) background class, 3) to-be-determined (TBD) region, obtained from the respective means of the other two classes. Iteratively the Otsuís method is applied onto the TBD region until the preset criteria is met keeping the foreground and background regions unchanged and finally their respective unions are formed for final result. The proposed work is to use this method and another method called Window based thresholding in order to reveal the fine structures in the Printed Circuit Board (PCB) in a minimal computational cost and achieve better accuracy for the same.
Keywords: Triclass, Printed Circuited Boards (PCB), segmentation, Threshold, Otsu's method, Window based thresholding, Binarization.