The Method And Realization Of High-Precision Size Measurement Based On Machine Vision

In recent years, with the prosperity and development of the auto market, auto parts processing and manufacturing also ushered in the golden age of its own development. Compared with the past30years, modern automobile manufactures raise higher demand on quality and machining accuracy of parts. Which not only requires enterprises to equip effective automatic production line to shorten its production cycle, but needs to innovate their measuring technology. For traditional manual measurement, complexities of monitoring in each process and unavoidable manual measuring error have long been a perplexity in quality control. To solve this problem, using visual measuring instrument for non-contact measurement has gained more popularity from mechanical processing companies.In order to improve measuring precision and efficacy, this paper takes the size measurement of automobile engine valve as research object. And research method and realization of high-precision size measurement based on machine vision. Measuring system mentioned in this paper uses high-resolution lined array CCD to acquire gray image of measured valve, eliminates noise with image preprocessing technology, locates measuring position with sub pixel precision, and calculates measuring result combined radial distortion correction algorithm and calibration parameters. The main research contents are as follow:(1) Classify valve size according to geometric characteristics, and design different valve size measuring methods to improve its generality and extendibility. Which will facilitate measuring new valve size.(2) To meet the requirements of high-precision size measurement, this paper analysis main factors to valve size measurement, which contains measuring environment, camera resolution, lens distortion and error processing. According to its formulation, we provide corresponding solutions.(3) Research method and realization of size measurement under standard parts, edge detection and light source. The paper points out how to design and manufacture standard parts, and using multiple direction Prewitt operator to detect different shapes edge. Finally, considering impact of changeable intensity on measurement precision, we put forward establishing mathematic model between light source intensity and pixel equivalent-value based on polynomial approximation. Experiments results show that this method is effective.