Three Dimensional Reconstruction Of Surfaces Based On Digital Grating Projection

At present,the method of using phase shift measurement contouring to obtain the surface morphological characteristics of objects has a broad application prospect in the fields of automotive,information security,aerospace,and robotics due to its convenience and accuracy.This paper takes the method of surface three-dimensional reconstruction based on digital fringe projection as the main line,and analyzes the existing three-dimensional reconstruction methods and application prospects,and studies the three-dimensional reconstruction technology method of digital grating projection from both theoretical and experimental practice aspects.The main research contents are: establishing the mathematical model of the measurement method,system calibration method,grating coding and decoding technology,grating image denoising,camera and projector calibration method,etc.Firstly,the background and significance of the research of this paper are explained,and the research status of the measurement system model and denoising method of streak grating image in digital fringe projection technology is analyzed,and the mathematical model and calibration method of camera and projector are studied according to the measurement system model.Secondly,the structured light encoding and decoding technology method is studied,the principle of phase shift method and time and space phase expansion algorithm are analyzed,and the multi-frequency heterodyne principle is selected for the feasibility analysis experiment.In order to obtain high-precision phase distribution,a combination of multi-level wavelet decomposition and neural network is proposed for stripe raster image denoising for the noise problem existing in the experimental raster images.The problem of raster image denoising is converted into the problem of predicting the transform coefficients in the wavelet domain,and the number of wavelet decomposition levels is determined and a new wavelet loss function is defined according to the analysis,which effectively "extracts" the potential noise information.Finally,an improved method for solving the system parameters is proposed to deal with the limitations of the traditional measurement model.The method uses the spatial coordinate conversion relationship to derive the system parameter solution method,and the calibration data obtained is employed to adjust the spatial position of the projector and camera in real time to complete the parallelism and perpendicularity calibration of the system,and improves the measurement accuracy of the system.Through the established experimental platform,the three-dimensional measurements of ping-pong balls,face models and rectangular hexahedra were completed,and the experimental results proved the effectiveness and availability of the method.