Fuzzy reasoning for pcb inspection

Using the concept of fuzzy reasoning, we developed an automated visual inspection system to detect defects of PCB (printed circuit boards) inner layer. The algorithm consists of two major processing phases. At the first phase, preprocessing is performed to eliminate noise and to reduce the number of potential candidates of PCB defects. At the second phase, features of PCB defects are obtained by image subtraction and defect boundary analysis processes. The results of the proposed fuzzy linguistic method are compared to well-known noise elimination algorithms. They are found to be of comparable quality with less computational complexity.