A new machine-vision-based approach for rapid dimensional inspection of an integrated circuit carrier tape

Abstract

This study presents a rapid and reliable machine vision technique for measuring the principal features of interest in an integrated circuit carrier tape, namely the diameters of the circular sprocket perforations and centre hole, the width of the carrier tape, and the width and length of the centre cavity. In performing the measurement process, the quality of the image acquisition process is enhanced by using two auxiliary light sources to suppress the effects of natural variations in the environmental lighting conditions. Having acquired the image using a charge coupled device (CCD) camera, the features of interest are separated from the background region of the image using a two-threshold algorithm based on the Otsu threshold selection method. The edge of each feature is then extracted from the binary image using the Canny edge detection method. The dimensions of the circular features are obtained by fitting four right-angle triangles within the periphery of the extracted circular edge and then computing the circle diameter by taking the mean of the hypotenuse values of the four triangles as computed using the Pythagorean theorem.