Method And Technology For Robotic Measurement Of Complex Surface On Large-scale Components

The large-scale equipment manufacturing plays an important role in basic and military industry.The quality of final product mainly depends on the surface quality of critical components.It is difficult to fabricate the surfaces of large-scale components because of their complex surface models.The shape measurement method has become the bottleneck of large-scale equipments' development and its technological progress in manufacturing.As an in-situ quality control process,the measurement method should satisfy the following conditions at the same time: large-scale,high density,high accuracy and high efficiency.In this dissertation,a research of automatic precision measurement method is proposed to satisfy the measurement demand of the complex surface on large-scale components.Based on the detailed review of the large-scale precision measurement methods and high density point cloud measurement methods,a combined measurement method,which integrates the advantages of fringe projection measurement,close-range photogrammetry and industrial robot,is implemented.This research is mainly focused on the improvement in the measurement accuracy of local point cloud and the automatic level of high accuracy point cloud registration.The experimental verification of complex surface measurement on large-scale components is completed with the rivet flushness inspection of aircraft skin and shape measurement of outer covering parts of car.The main works are listed as follows:1.A measurement method of complex surface on large-scale is proposed: The fringe projection measurement obtains high accuracy and high density point cloud in small scale.The close-range photogrammetry provides a high accuracy global control network to guarantee the accuracy of point cloud registration.The industrial robot implements flexible and automatic measurement.2.A measurement method focus on complex surface shape is implemented: to obtain the 3D point cloud,a stereo vision measurement system is built up with fringe projection technique.The four-step phase shifting method and three frequency unwrapped method are used to calculate the absolute phase.A fast matching method based on the stereo relationship of camera and projector is proposed to speed up the point cloud generation.The automatic matching method of feature points is also proposed for in-situ precise stereo calibration.A flexible precise projector calibration method based on stereo vision and a per-pixel full-field distortion correction method are proposed.The accuracy optimization method on fringe projection system,which including the pose optimization of cameras and the quality optimization of projected fringe patterns,is discussed.3.Global accuracy control is implemented based on improved close-range photogrammetry technique.The global accuracy control network is established with global feature points which are both obtained by fringe projection measurement subsystem and the close-range photogrammetry sub-system.The improved close-range photogrammetry technique is proposed with establishing the bundle adjustment equations and the control points coordinate transformation equations.The highaccuracy point cloud registration can be achieved with the help of the global accuracy control network.To further improve the extraction accuracy of the feature points,a precise concentric circles extraction method and the corresponding optimized feature designing method are also proposed.4.The automatic measurement is implemented with industrial robot.A hand-eye calibration method is proposed with a parameters optimization for the robot kinematics model and hand-eye relationship.With the robot kinematics model and the hand-eye relationship,the initial value for the close-range photogrammetry can be obtained quickly and reliably.The automatic measurement can be implemented by off-line designing and optimizing robot poses with the digital model of the large-scale components.