Research On Three-dimensional Measurement System Of Grinding Wheel Profile Based On Structured-light Vision

With the rapid development of artificial intelligence, the grinding process of machine parts which have complex curved surface like turbine blade have gradually moved toward automation due to the widespread application of industrial robots. However, in actual grinding process, not only the accuracy of robot path planning, also the feature parameters of the grinding wheel profile have a direct impact on the surface quality of workpiece. For example, full consideration should be given to deal with the radial weal of grinding wheel in order to take real-time compensation in robot path planning. The interference between the grinding wheel and the surface of wo rkpiece could be avoided through obtaining the curvature and the profile tolerance of the grinding wheel online. However, few attention have been paid to the change of characteristic parameters such as outer diameter, curvature and profile tolerance when g rinding wheel is in working state, which lead to the lack of measurement system for getting the contour parameters of grinding wheel. Moreover, due to the complexity structure, the available systems for profile measurement will make the wasting of resources, increase the cost of manufacture, and also the detection efficiency is low. In this paper, a three-dimensional(3D) measurement system of grinding wheel profile based on structured-light vision is designed. The feasibility of measurement system is valid ated by the experiment of measuring the grinding wheel profile. The system can be applied to obtain the contour parameters quickly in automatic grinding process.In this paper, the efficient detection method of central line of structured laser light stripe is studied firstly. Based on preprocessing the original image of light stripe, a rapid extraction algorithm of linear light stripe is proposed in view of the linear light stripe in target images when calibrating the light plane. For the central line of the light stripe of grinding wheel contour, The Gaussian fitting method is improved due to the random variation of curvature which ensured the extracting accuracy and velocity reliably. The calibration method of grinding wheel profile measurement system is studied secondly. Based on the principle of Zhengyou Zhang calibration, a nonlinear camera imaging model is established which achieved a high precision calibration to the geometry parameters of camera. The simple and efficient calibration of light plane is attained when adopting the extraction method of feature points in light plane based on planar target. According to the characteristic of the objective table rotating around the rotary center, a calibration method of rotary center is designed based on the chessboard target. The rationality of the calibration method is demonstrated through the calibration experiment of the whole measurement system. Then a 3D reconstruction model of grinding wheel contour is obtained. A curvature estimation method based on 3D scattered points is gotten by using coordinate transformation and constructing local paraboloid. The surface fitting of the contour of cylindrical grinding wheel and the conical grinding wheel is achieved based on the fitting method of 3D points. Finally, the feasibility of measurement system for measuring the contour of grinding wheel is verified according to the actual measurement results of the contour parameters.