Study Of Calibration Methods For Linear Laser 3D Measurement System

With the introduction of Made in China 2025,vision measurement has become one of the key technologies in the field of intelligent manufacturing due to its advantages of non-contact,high accuracy and automation.Line laser 3D measurement is one of the vision measurement technologies,which is widely used in the fields of intelligent manufacturing,reverse engineering and unmanned vehicles due to its better accuracy and real-time performance.The measurement system needs to use image processing technology to map the image plane to the three-dimensional space,and to obtain the mapping relationship between the two is called calibration.Therefore,the calibration accuracy will directly affect the system measurement accuracy and determine the value of the system usage.The paper focuses on the calibration theory in the line laser 3D measurement system,and analyzes the key techniques involved in the calibration process and the factors affecting the accuracy,so as to improve the accuracy of the system calibration and the system measurement accuracy.The main research of the thesis was as follows.(1)Measurement system equipment selection: The principle of laser triangulation measurement is analyzed,and the influence of the system structure parameters on the measurement range is verified by simulation experiments.According to the design requirements,suitable system hardware is selected,and the measurement range and resolution of the system are verified according to the system structure parameters.Finally,the scanned object model and calibration scheme of the measurement system are given.(2)Camera calibration: Analyzing the image quality evaluation method and the camera calibration method of Zhengyou Zhang,an automatic detection and ranking algorithm based on tessellation grid corner points is proposed for the problem of low robustness of Open CV corner point detection algorithm to illumination changes.And the calibration accuracy of the two methods is evaluated using the reprojection error after camera calibration,and the corner point extraction accuracy is improved by 61.87% compared with the Open CV library.Finally,the camera calibration experiment is completed using the detected corner points.(3)Laser plane calibration: The commonly used light bar centerline extraction algorithm is analyzed,and an improved Steger centerline extraction algorithm based on the region of interest is proposed for the problems of background interference and too wide light bars in images.For the problems of few feature points and slow calibration process of optical plane calibration,a calibration method based on the extended single response matrix and Plücker matrix is proposed to complete the laser plane calibration experiment.(4)Accuracy analysis of the measurement system: Repeatability measurement experiments using standard gauge blocks and round hole workpieces are used to verify the measurement accuracy and feasibility of the system.Also,the error sources of the measurement system are analyzed and the influence of such factors on the calibration parameters is verified using experiments for the human factors in the collected calibration images.Finally,the measurement accuracy and repeatability accuracy of the system are verified through the debugging of the combined software and hardware.Among them,the root mean square error and sample standard deviation were used to evaluate the accuracy and repeatability,respectively.From the experimental results,it can be seen that the system has high measurement accuracy and good repeatability,both of which are within ±0.1mm.