Abstract
The inspection of wave-soldered joints is a critical step in the assembly of printed circuit boards (PCBs) of high reliability. The vast majority of this inspection is presently performed by humans, however studies have shown that the inspection results are not reliable as they vary with mood, time, experience, and personal interpretation skills. Research efforts as reported in this paper are directed at the automation of this solder joint inspection through the use of machine vision, artificial intelligence, and neural networks. It was determined that defective wave-soldered joints can be distinguished by using the characteristic shapes of the image grey-scale histograms. The proposed method of intelligent histogram regrading (IHR) incorporates a modified version of a previously reported adaptive histogram regrading (AHR) technique that divides the histogram of the captured image into different modes. Each distinct mode is identified and the range of grey levels corresponding to each mode is separated and regraded using neural networks. Feature selection is performed by neural networks in terms of certain variables (values) for the first time. The inference engine then uses these values and supplied rules to identify and classify the defective solder joints. A system based on this IHR method is outlined and several performance evaluation parameters are presented.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call