Rotation Correction Method of Thread Image Based on Machine Vision

摘要

In order to obtain accurate diameter parameters when measuring cylindrical thread by machine vision, it is necessary to calibrate the measuring coordinate system of acquired thread image. Firstly, the principle, model of thread image rotation correction is established. Secondly, the relationship between the deflection Angle of the thread image and the measurement result of the middle diameter of the thread is obtained by theoretical pushover. An algorithm flow of image rotation correction is proposed. At last, the thread plug gauge is taken as the experimental object to carry out the comparative test. The experimental data show that the proposed algorithm can realize the correction of thread image, the measurement data fluctuated in a small range, and the measured diameter variance of thread is reduced by 77.1 % compared with the traditional correction algorithm. It is shown that the proposed correction algorithm can accurately and efficiently measure the parameters of skew thread images.