The Development Of Visual Inspection System For Surface Mount Led Automatic Taping Machine

Taping machine is a automation equipment that bring semiconductor chips into the carrier to facilitate subsequent equipment deal with semiconductor chips. In order to ensure semiconductor chips quality in the carrier tape, precision visual detection system of tape machine has a faster speed and greater accuracy to capture a variety of defects, and automatic classification, theory and analysis of new technologies can be used to detect various defects. Therefore, visual detection system is very important part of tape machine.Traditional visual inspection system for automatic taping machine is similar to general inspection system of foreign, but it has a lot of problems such as high costs, lack of core technology, operations complex and so on. Therefore, developing independent intellectual property rights of surface mount LED visual inspection systems for automatic taping machine has very important practical significance. This article has designed a visual inspection system of taping machine. First, this article has completed the requirement analysis of LED, and determined on the visual inspection system’s the work to be completed, and finished designing the visual inspection system of taping machine’s general structure. Second, completed hardware selection of visual detection system of taping machine according to the demand analysis, mainly containing CCD camera, image collection card and a camera lens. Finally, According to visual inspection system of the taping machine’s the work to be completed, determine and implement detection of the various project’s algorithm.This system has mainly completed detection of direction and missing of LED. First, deal with the image preprocessing before detection of direction and missing of LED. Image preprocessing is mainly to want to eliminate outside interference with the image, in order to achieve the best results. Taping machine is subjected to salt and pepper noise and impulse noise interference in working process. Following comparative experiment, this article adopted median filtering algorithm to do with image preprocessing. Second, position correction of image method is to design correcting window, make position migration data of correction window reflected in other detection range automaticly. Next, about detection of direction and missing of LED, the detection of direction’s algorithm is to calculate the area of image after binary image, count white pixels or black pixel’s area, and compare deviation of the area of the different Angle to find position of missing angular, so as to determine the direction of LED. Missing detection’s algorithm is to calculate brightness value of image, and compare deviation of brightness value to judge whether there is material in carrier tape. Finally, this article puts forward the way of segmenting image of0603green’s LED by MeanShift, and then extract the edge of image, last analyse pin number of image internal to judge the direction of chips.