A Design Of Measuring System For Female Thread Parameter Based On Machine Vision

In the industrial production and processing, the role of the nut is self-evident.There are different requirements about the quality of the nut parameters in different operating environment, so the measurement method and efficiency of nut female thread is not the same according to different requirements.Aiming at the faults in the female thread detection, such as, poor efficiency, low precision, high cost,damage caused by contact measurement and so on,a non-contact and efficient measure method based on machine vision is proposed in the thesis,and a measuring system was designed. According to the nut features to be measured, a method of image acquisition is designed, which can acquire high quality images of female thread quickly, and the symmetry of the image acquisition and measurement method greatly reduces the measurement error due to eccentric nut. And, according to the features of the images acquired,heightening contrast ratios of images by taking a method that gray images selectively, and weakening effect of scattering from surface,conduct the subpixel edge detection for image after filter, and mending cracks by using the Close Operation,get the piecewise interval, then conduct the least squares polynomial fitting for subpixel edge points in each interval. The fitting curve regarded as female thread contour line in which positioning the female thread feature points, the coordinates of feature points is used to calculate the related parameters of the female thread.The system based on the method above features the precision of 12.78?m for thread internal diameter repeated measurement, and average time no more than 200ms from image acquisition to the end of the parameter measurement.