Pattern defect analysis and evaluation of printed circuit boards using CAD data

A methodology for optical inspection of large printed circuit boards using their original artwork data or CAD data is presented. The manufacturing process and some ambient and control conditions are evaluated directly from the optical image using the absolute data. An accurate automatic inspection system can be also expected. The experimental evaluation shows that an optimal size of the allowable skeletons margin is about 40 to 50% of the average width of the conductor patterns. The inspecting skeletons should be sampled at variable intervals, rather than taking all the pixels of the skeletons, so that the accuracy and the comparison time may be compromised. These preliminary experiments are essential for the practical development of the system. It is demonstrated how the current system is improved from conventional ones in many aspects.<>