Research On 3D Surface Topography Measurement Technology Of Objects In Complex Environment Based On Polarization Information

Fringe projection profile measurement(FPP),as one of the popular technologies of non-contact three-dimensional(3D)sensing/imaging,has been rapidly developed in the past few decades.3D optical sensing has begun to become an indispensable part of our daily lives,such as devices such as Face ID enabled by 3D sensors on smartphones.With the continuous application of FPP in artificial intelligence(AI),machine learning,intelligent manufacturing,robotics and other fields,such technologies will have a more profound impact.However,this advanced optical system still has some basic problems in industrial applications.In the field of structured light three-dimensional measurement,the measurement accuracy of the system is an important index to evaluate the overall performance of the system.Among them,the accuracy of system calibration directly plays a decisive role in the entire system.Therefore,how to accurately perform system calibration has always been an important content of structured light projection measurement.At present,the system calibration method based on the plane target is widely used because of its simple principle and easy operation.However,because the projector is a strong light projection device,when the projected target image is collected by the camera,the pixels in the edge area of the target image often present a strong saturation state.These saturated pixels affect the extraction of the sub-pixel coordinates of the feature points,resulting in lower calibration accuracy of the projector.Aiming at such problems,this paper starts from the angle of light polarization,and proposes to collect the target image under the optimal polarization angle,eliminate part of the highlight area of the image,and restore the texture detail information of the corner edge.At the same time,a sub-pixel corner extraction algorithm based on homography matrix mapping is proposed to overcome the problem of incorrect or missing corner detection in the traditional Harris corner extraction algorithm,and improve the accuracy of corner extraction.In terms of system parameter optimization,adaptive weights and mutation particle swarm algorithm are used to perform nonlinear optimization of camera and projector parameters.Through the individual search ability,the inertia weight is dynamically adjusted,and the mutation operation is introduced,which overcomes the shortcomings of traditional optimization algorithms and improves the global search ability,so that the system can be calibrated quickly and accurately.On the other hand,in FPP,a series of sinusoidal fringe patterns are usually used to obtain phase information for the restoration of the three-dimensional geometric shape of the measured object.In industrial production,the measurement environment is usually more complicated.In addition to the projected light source of the system,there are many interfering light sources in the environment.These interfering light sources will cause certain interference to the projected fringe image during measurement,causing the intensity of the captured fringe image to deviate from the true value,resulting in errors in the decoding of the image.Regarding the non-linear effects of different environmental light sources on the measurement system,this paper proposes a new structured light 3D measurement method that combines phase and polarization.Regarding the influence of the ambient light intensity on the quality of the fringe during the measurement,the Stokes parameter is used to indicate the polarization state of the ambient light after being reflected by the target object.The theory proves that the impact of ambient light on the system is only related to the Muller matrix of the target surface.Use Fourier transform to enhance fringe contrast through frequency domain filtering.In addition,use an improved spin filtering method for image edge denoising to effectively remove the impact of ambient light sources on the system,and finally obtain high-quality phase-shifted images Recover the three-dimensional outline of the target object.Finally,a set of structured light three-dimensional measurement system for diffuse reflection objects is developed to perform three-dimensional contour measurement on the surface of objects with different reflectivities in complex environments to verify the effectiveness of the proposed algorithm.