Research On Handheld Structured Light Measurement Technology Based On Binocular Vision

At present,three-dimensional measurement technology is widely used in high-speed trains,aerospace and other precision manufacturing fields,in which handheld structural light measurement technology is welcomed because of the advantages of flexible operation,wide measurement range,measurement accuracy and so on.However,due to foreign monopoly of technology and other reasons,the domestic research on this technology started late,the development is slow,and for the box,frame and other large depth measured object measurement results are poor,it is difficult to achieve the requirements of high-precision three-dimensional measurement.Therefore,this paper studies the handheld structured light measurement technology based on binocular vision as follows.(1)A handheld structured light measurement system is constructed.The system consists of a binocular camera,a planar target,and a structured light sensor,etc.The binocular camera is used to solve the positional solution of the rigidly connected planar target and the structured light sensor,so that the local contour data of the object measured by the structured light sensor can be transformed to a unified world coordinate system,laying the foundation for efficient measurement work.(2)A high-precision visual imaging model for multi-depth-of-field targets and its calibration method are proposed.The method is solved to obtain the true values of camera parameters at different depths and applied to actual measurement activities,which effectively solves the problem that it is difficult to further improve the visual measurement accuracy in the field of high-precision large-depth 3D measurement,and designs verification experiments to prove the algorithm effect.(3)The corresponding image processing methods are designed according to the image characteristics of planar targets and structured light bars.Firstly,we use grayscale transformation and image filtering methods for pre-processing to remove noise interference,and secondly,based on the analysis work of traditional algorithms for target detection,circular marker point contour and center coordinate extraction,YOLO neural network and Hough circle transformation algorithms are integrated and selected to extract effective information.(4)The systematic measurement method is studied to establish the camera coordinate system,the plane target coordinate system,and the structured light plane coordinate system and solve the transformation relationship between any two one by one.Among them,after analyzing the classical structured light calibration method,a modified tab-type stereo target is proposed for solving the rigid body transformation matrix between the camera coordinate system-structured light plane coordinate system.Finally,the joint calibration results are applied to the actual 3D measurement activities,and the reliability and superiority of this system are proved by the accuracy analysis of the measurement results.The experimental results show that the accuracy of solving the attitude of the planar target with the parameters obtained by the camera calibration method proposed in this paper is nearly 1 times better than that of the classical camera calibration method.The image processing method proposed in this paper can quickly and accurately extract the ROI area,circular marker point contour and center coordinates.When the rigid body transformation matrix solved by the joint calibration of the system is applied to the actual 3D measurement activities,the average accuracy of multiple measurements can reach 0.03 mm,which meets the industrial measurement requirements.Overall,the system can measure the whole object quickly,accurately and efficiently.