Innovative Image Segmentation Method Employing Fuzzy Logic Algorithm

Liang-Chia Chen and Xuan-Loc Nguyen


Automatic Optical Inspection, Print Circuit Board, Fuzzy Logic, Object Segmentation


This article presents a new method for image segmentation using fuzzy logic algorithm to overcome the existing difficulties encountered by complicated intensity variation and low contrast conditions. Effective object segmentation is critically essential to component detection and positioning in automatic optical inspection (AOI). The proposed method employs the relationship of the intensity of a pixel and its neighbors to segment the image underlying inspection into two clusters. By defining a set of critical points for the training process in fuzzy-logic rule base, the method is capable of extracting the detecting object from a low-contrast and noisy background image. By defining an adequate set of the segmentation parameters in the method, it is proved to be effective and robust for component segmentation in an image background having various conditions of lighting, view angles and detecting areas. In addition, from the experimental results, it was found that the performances of the developed method is controlled by its model parameters such as the number of learning points and the mask size for intensity calculation.

Important Links:

Go Back