| Phase-shifting fringe projection is a non-contact 3D measurement method,which can obtain 3D data of the measured object quickly and accurately.This method can obtain accurate measurements for the measured object without specular reflection.However,when the measured object contains specular reflection,saturated pixels are generated in the images collected by the camera.Highlights make the phase calculation error,resulting in a large deviation between the measurement results and the actual data.Therefore,this paper designed and built a 3D measurement platform for highly reflective objects based on phase-shifting fringe projection,and studied its related technologies,such as calibration of measurement system,Gamma distortion correction and 3D measurement methods for highly reflective objects.The main work contents and innovations of this paper are as follows:1.To address the difficulty of measuring high light surfaces,a 3D measurement method of highlight surface based on deep learning framework is proposed.DC-Unet and DRN networks are used to repair the highlight in fringe image.Firstly,the simulated experiment is implemented to verify the effectiveness of the proposed method,and then the real 3D measurement experiments of objects with highlights are implemented with the same method.Experiment results show that the proposed method can reduce the measurement deviation caused by the highlights.2.To address the difficulty of measuring high light surfaces,a 3D measurement method of highlight surface based on adaptive fringe pattern projection is proposed,which reduces the specular reflection component of the measured object surface and avoids the highlight in the camera image by reducing the projection intensity of the projector in the highlight area.The measurement experiments were implemented on the measured objects made of metal,plastic,and ceramic,respectively.The experimental results indicate that the proposed method can reduce the holes caused by highlights and make the 3D reconstruction model more complete without additional cameras and hardware.3.In order to solve the problem of difficult to extract the corner phase of the calibration plate in the calibration process of 3D measuring system,a method for evaluating the phase quality of corner points is proposed in this paper.This method can calculate the phase quality coefficient of corner points by comparing the phase of corner points and its surrounding phases.The experimental results show that the calibration accuracy of the system can be improved by choosing the corner phase with higher quality coefficient.4.In order to quickly generate and collect fringe images,Qt and Open CV are used to write the corresponding software for the measurement system.This software can recognize the resolution of projector,automatically generate sinusoidal fringe image sequence with specified parameters,control the projection of fringe image by projector and control the acquisition of image by camera.In order to improve the measurement accuracy,the function of Gamma distortion correction is added in the software,which can generate the fringe image with Gamma pre-correction. |