Research On Specular Shape Detection Technology Based On Fringe Reflection

With the development of modern optical industry,aspheric mirrors have been widely used in advanced manufacturing,aerospace and other civil or military fields for their unique advantages such as correcting system aberrations.The surface accuracy of aspheric mirrors directly affects the performance of the system,but the traditional detection methods have some limitations.Fringe reflection detection technology,with its advantages of high accuracy,high efficiency and non-contact,is of great significance in the measurement of three-dimensional topography of aspheric mirrors.This paper discusses the system of specular shape detection based on fringe reflection and the key technologies involved in the system,to realize the noncontact measurement of aspheric mirror.The paper first expounds the principle of fringe reflection detection technology,the monocular measurement model based on plane iteration and the binocular measurement model based on normal matching.In the monocular measurement model,it provides the mirror height by using the plane for iterative calculation which do not need to provide a prior information of the test mirror.In the binocular measurement model,the surface gradient of the mirror to be measured by matching the normal direction of the two cameras.In terms of the discussion of the key technologies,the phase calculation formula of phase-shifting algorithm in digital phaseshifting technology is deduced.And the phase resolution accuracy of two global phase unwrapping technologies,binary coding and dual-frequency coding,is compared and analyzed under the condition of random noise and defocusing error.Then the Fourier transform integration of surface reconstruction technology based on gradient data as well as the reconstruction effect after introducing smoothing constraints is analyzed.In the calibration process,a plane mirror with the same thickness as the calibration plate is used to calibrate the display's space position,which realizes the unification of camera and display coordinates in the world coordinate system where the calibration plate is located.What follows is the analysis of the effect of the calibration board accuracy on the calibration results,as well as the corner detection accuracy of three calibration board image corner detection algorithms with random noise and defocusing error.Then,the principle of the two measurement models is verified by simulation.On this basis,two topics are analyzed,including the influence of the initial iteration plane at different heights on the shape measurement results and the convergence ability of the monocular measurement model,as well as the influence of the quantization of matching data on the shape measurement results in the binocular measurement model.Finally,a specular shape detection system based on fringe reflection is built.In addition,the corresponding calibration plate and plane mirror are made to complete the calibration of the detection system.The influence of lens defocusing on Gamma effect of display during measurement is analyzed,and the global phase resolution accuracy is tested.Furthermore,the two measurement models are applied to the detection of 80 mm parabolic mirror.The results indicates that the detection accuracy is close to the micron.