Research On 3D Projecting Indication Technique Based On Binocular Vision

With the continuous improvement of product quality and producing efficiency,the manufacturing industry has put forward new requirements for the detection technology in the production process.Aiming at the problem of pipeline docking deviation during the assembling process of large ships,the existing detection method is that operator manually detects and marks the surface of the workpiece which is inefficient and cannot guarantee measurement accuracy.In this paper,a three-dimensional projecting indication system based on binocular vision is studied.The binocular camera is used to locate the feature points in the three-dimensional space,and then the projection device is used to project the measurement results onto the surface of the workpiece being measured,which realizes measurement visualization and improves the digitization and automation level of the assembly and inspection process.The non-contact measurement method effectively reduces the wear on the surface of the workpiece,and projecting indication simplifies the process of reading measurement data through the terminal display equipment which improves the flexibility of the detection.On the basis of realizing the measurement indication,the method of regional pixel reconstruction based on perspective transform and projector parameter calibration based on the bundle adjustment are studied respectively to eliminate the distortion of the text information and improve the projecting accuracy,which optimized the effect of projecting indication.Through the experiment,the circular reflection point on the outer cover of an automobile is measured and marked.The results show that the positioning error of the system in global coordinate system is 0.1-0.4mm,the measurement result can be clearly displayed on the workpiece and the indication error does not exceed 1mm.Overall,the technique has high practical value.The main reaserch work are as follows:1.The mathematical model of the 3D projecting indication system based on binocular vision is studied.While the projecting indication of the measurement result is completed,the distortion of the auxiliary text information caused by the position of the projector is also analyzed.And the text information which cannot be clearly displayed is reconstructed by pixel in the indication picture.2.The calibration requirements in the projecting indication system are analyzed.The calibration of the camera and projector is completed based on Zhang's calibration method and back projection method,and the calibration accuracy is analyzed.The internal and external parameter calibration principle and distortion rectification process of binocular camera system and camera-projector system are introduced respectively.3.The key techniques of binocular visual image analysis and processing are studied.An optimization algorithm based on Hough transforms and the polar constraint method are used to extract and match the feature points on the surface of the workpiece.The error between the projecting indication position and the actual position of each feature point is analyzed.Based on bundle adjustment,the indication accuracy is improved.And the parameters of the projector are optimized by the Levenberg-Marquadt algorithm.4.A three-dimensional projecting indication system based on binocular camera is built.An experiment is carried to verify and analyze the accuracy of the methods of system calibration,binocular vision measurement,projecting indication,regional pixel reconstruction and projecting indication accuracy optimization.The experimental results show that the method is effective and can be flexibly applied in the process of detection.