Automatic surface mount solder joints inspection

This thesis reports the research results on automatic inspection of solder joints on printed circuit boards. The previous work on this subject has been advanced significantly in the following three aspects. With the support of AT&T Bell Lab, the most updated surface mount solder joints are inspected in this work instead of larger simulation solder joints or traditional through hole solder joints in the previous work. A small set of features is extracted for surface mount solder joints in both infrared and visual light inspection. A new image processing software named Khoros has been applied to improve the quality of images. It has been demonstrated that infrared imaging technique can identify solder joints of surface mount printed circuit boards according to their solder volumn. The correct classification rate was found to be in the range of 89% to 100%. For the sample joints provided by AT&T Bell Laboratory, reasonably good inspection results have been obtained. The experimental results demonstrate that infrared imaging technique can be utilized to discriminate solder joints on surface mount printed circuit boards with different solder volumes quite reliably