Summary: | 碩士 === 國立屏東科技大學 === 機械工程系 === 91 === In this paper, an inspection system including the machine vision and a inspection mechanism is self-designed to inspect the size appearance of SOP IC component. The inspected items are coplanarity, pitch, sweep, stand off and spread etc.
The proposed system is collection of twofold subsystems. It consists of the inspection mechanism and inspection software. As for the inspection mechanism, the mechanism can be divided into the stations of loading, separation, alignment, inspection, rotation and output sorting. In inspection software phase, to attain the robust results in the repeatability and accuracy of the inspection, the image processing methods including binary-thresholded, blob-analyzed, edge-detected and subpixel technique are employed in the present thesis. The results of the inspection repeatability from a series of experiments show that the level of GR&R is within ten percent. Meanwhile, the measured results of comparing a projector instrument and the present inspection system demonstrate that the maximum deviation is not greater that 0.48mil for all of the measured samples.
From the overall view of evaluation on the present developed system, the repeatability and accuracy of the inspection ability with the function of automation system can effectively replace the bared eyes to meet the requirement of industrial application.
keywords:coplanarity、pitch、sweep、stand off、spread、blob analyzed、edge-detected、subpixel analgsis、repeatability accuracy。
|