The machine vision is a sort of machine which simulates the human eyes' function. It extracts information from the image or image sequences andrecognizes the form and movement of the 3-D scenery. It's a noveltyresearch field which develops rapidly. From the birth, it has fifteen yearshistory. And it will go ahead with develop of the industry automatization.Now the whole market of machine vision is about sixty hundred milliondollars, which increases 8.8 percent every year. The advanced technology isbased on the image-taking device and the machine vision theory, whichrealizes the industry inspection by using image processing, image analysisand module recognition with fewer investment, better adaptability and easymaintenance. The machine vision does an important role in many fields,such as industry, medicine, remote sensing, identification, surveillance.The coaxial percent computing of micro-component has much to dowith some knowledge, such as pattern identifying, image processing,machine vision, mechanics and so on. It will be a big market in apply ofmanufacture.This paper point out an algorithm of computing coaxial percent basedon digital image processing theory. The biggest difficulty is the electroniccomponents are made in the same material and have the same color. Theonly method to distill brim is based on the different luminance of the imagebecause of the scraggly superficial. In low contrast condition, there is littleimage noise, but the inside brim is faint. In high contrast condition, theinside brim is clear, but there is much noise. The goal of this paper is tobuild up a system, which is less influenced by the environment.In the first part, this paper introduces the frame, contents andforeground of machine vision. Based on digital image processing, a methodis mentioned about how to computer coaxial percent of micro-bugle. This method is an application of machine vision in the field of geometryparameter detection of micro-components and offsets the shortcoming ofold methods. The second part does some research of digital imageprocessing, preparing for finding an algorithm to distill the inside andoutside circles' brim. The third part is about the whole hardwareconformation, including optics system, image collection system and imageprocessing system. The last part is about how to computer the coaxialpercent and system errors.The research is completed in the Matlab platform, uses Matlab torealize all of image processing algorithm. Experiment result shows that themethod has better effect. |