Multi-view 3-D Measurement System Based On Composite Fringe Phase-shifting Profilometry And Five-axis Motion Stage

The rapid development 3C manufacturing industry has put forward a wide range of requirements for the three-dimensional inspection of meticulous and complex parts.The application of various non-contact 3-D optical inspection technology has become an important driving force for the development of the manufacturing industry.Among them,the structured light 3-D inspection method has been widely concerned with and studied because of its efficiency.At present,limited by the contradiction between a large field of view and high precision,the time-consuming 3-D algorithm,and the inflexible inspection process,the structured light 3-D inspection technology is also difficult to promote the application on a large scale.This thesis is oriented to the application scenarios of 3-D high-speed precision inspection in 3C manufacturing.Aiming at several key problems in structured light 3-D inspection technology,we deeply study the principle of 3-D imaging and the multi-view measurement method and propose a composite stripe pattern coding method,a triangular grid model construction algorithm,and a multi-view point cloud splicing algorithm.A 3-D measurement device and a five-axis motion stage were built,and a 3-D measurement software system was realized to verify all of the proposed methods in this thesis.The main research contents of this thesis are summarized as follows:(1)Through extensive and in-depth investigations of various domestic and foreign optical 3-D detection technologies,we summarize the research progress of structured light projected profiling in imaging principles and methods,analyze the key issues in detection accuracy and speed improvement,measurement range expands,and multiview measurement and then clarify the research content of this thesis.(2)Existing structured light 3-D inspection methods are easily affected by the projector's nonlinear light intensity response,resulting in nonlinear error.Aiming at this problem,this thesis proposes a new fringe pattern coding method,compound sinetrapezoidal phase-shifting coding.This coding can simultaneously code the measured object profile and the light intensity response model to realize the nonlinear intensity correction of the captured image,thereby effectively reducing the nonlinear error and improving the 3-D measurement accuracy without increasing the number of patterns.(3)Aiming at the point cloud noise reduction and geometric measurement technology in the structured light 3-D inspection system,the voxel filtering and down sampling methods of the point cloud and the 3-D geometry fitting algorithm are studied,and a modified sweep-circle algorithm is proposed to efficiently construct a triangular mesh.(4)In view of the inefficiency and accuracy of the multi-view point cloud splicing algorithm,a five-axis motion stage-assisted splicing method for large fields of view and multi-angle 3-D imaging is proposed.Through the kinematics modeling,stage calibration,and two-step registration algorithm base on stage pose information,realize effectively multi-view point cloud splicing,and verify the proposed algorithm through experiments.(5)On the basis of the above-mentioned research,a set of structured light 3-D inspection integrated systems based on a five-axis motion stage is developed.A 3-D inspection software system is implemented through C++,QT,Open CV,and Open GL library to verify the effectiveness of key algorithms of this thesis.