Research On Online Intelligent Measurement System Of Shaft Parts Based On Machine Vision

Shaft parts are one of the key components in the gearbox,and its machining accuracy will directly affect the transmission accuracy,performance and service life of the gearbox.Therefore,after the processing is completed,the shaft parts must be strictly measured according to the design requirements of the process file,and the subsequent process can only be carried out after qualified.Under the background of mass production,the online measurement system of shaft parts needs to have the characteristics of high precision,high efficiency,automation and intelligence.This paper focuses on the online measurement of the machining dimension of shaft parts,and develops the online intelligent measurement system of shaft parts based on machine vision technology.The main research contents of the thesis are as follows:According to the characteristics of the acquired part image,the morphological closed operation and bilateral filtering are used to preprocess the image,so as to ensure the continuity and integrity of the edge,and also eliminate the interference of the false edge caused by burrs or flocs on the detection accuracy.The OTSU method is used to adaptively determine the double threshold of Canny operator for edge detection of different images collected online,and the initial positioning of the integer pixel edge of the image of the shaft parts is completed by using its high robustness.Accordingly,the Steger sub-pixel edge detection based on Taylor polynomial fitting is carried out to solve the problem of large computation and low detection efficiency caused by traversing convolution on the whole image.The positioning accuracy of the edge detection algorithm is evaluated by the designed standard binary image and standard bar image,and the execution efficiency of the algorithm is statistically analyzed.By comparing with the detection results of Canny-Zernike moment algorithm and Canny-Franklin moment algorithm,the CannySteger algorithm proposed in this paper has high positioning accuracy and fast detection speed,which can effectively improve the edge detection accuracy and efficiency of the part image ROI,and quickly provide the geometric feature data point set of the measurement item,which has engineering application value.In order to improve the automation and intelligence of the online measurement process of shaft parts in the context of mass production,a comprehensive framework for online intelligent measurement of shaft parts is proposed based on the production process of parts.By analyzing the requirements of the measurement system,the overall architecture design of the measurement system is carried out.The measurement system structure,control system,and visual system are designed according to the measurement accuracy,rhythm,and other aspects of the measurement system.The measurement action process is determined to ensure the acquisition of high-quality images.Based on the DXF underlying data format of CAD process drawings,the process information analysis process of shaft parts is designed to extract the data of part contour and dimensional tolerance.The registration method of PCA analysis+NDT+improved ICP is proposed to realize the registration of image edge point set and DXF contour point set to determine the reference point of the measurement item in the image.The automatic measurement of geometric elements in the process can be realized by dividing the measurement area in the neighborhood of the reference point.After the calibration of the measurement system,experiments are conducted on the radial and axial measurement accuracy and repeatability accuracy.Through SPC analysis of measurement data during the production process of parts and the closed-loop system of machining,measuring parameter adjusting and reproducing,the quality control for the production process is achieved to ensure high-quality and continuous processing of parts.Based on the.NET platform,a measurement system software is developed using C#language to implement the hierarchical and modular design,which mainly includes control interaction,image acquisition and processing,information management,quality control and other modules.The developed online intelligent measurement system has theoretical and practical value for the automation,digitization,and intelligence of production lines to achieve large-scale,unmanned,and 100%measurement of shaft parts.