Research On Key Technologies Of Three-dimensional Reconstruction Of Monocular Structured Light Based On Fringe Projection

High-precision three-dimensional measurement of parts can effectively guide the adjustment of processing parameters,which is one of the effective ways to improve product quality and reduce production costs.Visual measurement technology has attracted much attention in the dimensional measurement of industrial parts because of its advantages of non-contact,high precision and speed.Among them,computer 3D visual inspection technology based on structured light coding is one of the mainstream technologies in the field of industrial 3D measurement,mainly including monocular structured light system and binocular structured light system.Compared with the binocular structured light 3D imaging system,the monocular structured light 3D imaging system only needs 1 camera and 1 projector,which has the advantages of compact structure and low cost.This question is primarily concerned with the key technologies of fringe projection monocular structured light three-dimensional imaging system,the specific content is as follows:1)A phase unfolding method with folding phase assist is proposed.Based on the positive and negative of the folded phase,the binarized image and the reverse binarized image are determined,and the connected domain coding function is used to obtain the connected domain code of the binarized image and the reverse binarized image,so as to obtain the fringe level.This method does not require additional image assistance to achieve phase unfolding,and has strong noise immunity,taking into account both fast and robust.In this academic dissertation,the technical links such as system calibration,nonlinear distortion and phase development in the three-dimensional imaging technology of stripe projection monocular structured light are studied,and the feasibility of the proposed scheme is verified from the perspectives of theoretical analysis,simulation analysis and experimental analysis,which has certain reference significance for personnel in related fields.2)In view of the significant influence of gamma distortion in three-step phase shift,a gamma distortion correction method based on least squares method is proposed.In this method,the target equation is established based on the minimization of plane fitting error,and the gamma distortion value is obtained based on the nonlinear optimization algorithm,which has the advantage of simple operation.Simulation and experiments verify the feasibility of the proposed scheme.3)Aiming at the influence of calibration plate pattern features on feature point phase extraction,a robust system calibration technique based on random sampling consistency(RANSAC)algorithm is proposed.The black area of the calibration plate pattern(such as checkerboard,black background and white background calibration plate)affects the sinusoidality of the captured stripe image,resulting in easy phase error of feature points.Therefore,based on the coplanarity of the phase of the feature points,the RANSAC algorithm is used as a plane constraint to filter out the feature points with phase noise,so as to improve the calibration accuracy of the system.This method is suitable for different types of two-dimensional calibration plates and has good robustness.