Research On Key Technologies Of 3D Reconstruction And Defect Detection For Aero-engine Blades Parts

Aero-engine blade parts are important parts in military and civil aero-engines.They have complex shapes and can only be sent abroad for repairs after damage,resulting in long maintenance periods and high costs.Therefore,speeding up the research and development of key technologies such as perception,measurement,repair,grinding and polishing in the repair process of aero engine blade parts is of great significance to the development of our country’s aviation industry.The 3D reconstruction and defect analysis of blade parts based on 3D vision technology is a key process and the first link in the aero-engine blade parts repair and grinding and polishing robot system.The 3D reconstruction accuracy,speed,defect recognition accuracy of blade parts and other indicators are key indicators of the robot repair and polishing system,which involves a series of key technologies such as system construction,system calibration,point cloud splicing,and defect detection.Based on this,this subject has conducted in-depth research on key technologies of 3D reconstruction and defect detection of aero-engine blade parts from four aspects: 3D structured light vision system design,3D vision system calibration method,3D reconstruction algorithm,and defect detection algorithm.The research results are promising,for providing theoretical basis and technical basis for subsequent automatic repair and polishing of robots.The main research as follows:(1)Through the research of camera imaging principle and projector imaging principle,the most widely used Zhang’s calibration method is used to complete the calibration of camera,projector,camera-projector,binocular camera and turntable hardware equipment in the 3D structured light vision experimental platform.Obtain the internal parameter matrix and the external parameter matrix of each device to provide basic parameters for subsequent 3D reconstruction.(2)Through the research of three-dimensional reconstruction principle and three-dimensional reconstruction formula,completed the three-dimensional reconstruction algorithm,and used the algorithm to carry out three-dimensional reconstruction experiments on real aero-engine blades.Based on the process flow of aero-engine blade parts repair and grinding and polishing robot system,this paper designs a 3D structured light vision system,which includes a hardware system and a software system.Through the design and selection of the equipment in the 3D camera module,the 2-axis turntable module and the three-coordinate robot hardware system,the design of the 3D structured light vision experimental platform is completed.Based on Visual Studio 2017 development environment,Qt、OpenCV and PCL third-party libraries,using C++ programming language,the functional modules of four software systems include system calibration,projector projection,3D reconstruction,point cloud registration and defect detection.The accuracy of the reconstruction system is verified by using a standard ball,and it is concluded that the system reconstruction deviation is about 30μm.(3)The rough stitching and fine stitching algorithms are studied,and the point cloud registration experiments are carried out under 4 viewing angles of the turntable and 2 binocular viewing angles to verify the reliability of the registration algorithm.By comparing traditional two-dimensional defect detection and three-dimensional defect detection methods,a defect detection method based on Euclidean distance is proposed,and different colors can be used to display the size of defects in three dimensions.The defect detection method is used to verify the pit defects and bump defects by experiments,which proves the reliability of the method.