Research On Key Technologies Of 3d Reconstruction Based On Spatially Coded Structured Light

Structured light three-dimensional reconstruction technology is a technology that uses optical means to encode the measured object,thereby acquiring three-dimensional data of the object.This technology has the advantages of high accuracy,fast speed,and good robustness.It is widely used in reverse engineering,biomedicine,cultural relic restoration,and face recognition.It has important theoretical significance and research value.With the rapid development of science and technology,the demand for accurate and efficient 3D reconstruction of targets has gradually increased,and high-precision 3D reconstruction of structured light has become a research focus.This paper mainly studies the three-dimensional reconstruction algorithm and its implementation based on spatially encoded structured light.The main work is as follows:(1)The structured light spatial coding technology is studied.Aiming at the problems that the traditional color fringe coding method is not suitable for complex-color objects and the fringe decoding is dependent,etc.,a structured light pattern composed of sinusoidal fringe pattern and pseudo-random dot pattern is proposed.The proposed pattern is a structured light pattern using two colors.The sinusoidal fringe pattern and the pseudo-random dot pattern are coded by two different color channels respectively.The fringes in the proposed pattern do not need to be identified by color information but are identified by the distribution of pseudo-random points around the fringes to avoid the interference of the color of the object on the color recognition of the fringes.Besides,the spatial similarity of the pseudo-random dot pattern is used to match the fringes,which improves the accuracy of the fringes matching.(2)The structured light decoding technology is studied.Aiming at the problem of sparse of 3D reconstruction data due to coding limitations in traditional structured light spatial coding methods,a phase-based structured light decoding algorithm with high spatial resolution is proposed.First,based on the fringe center points extraction algorithm,the fringe center points are accurately obtained.The spatial similarity of the pseudorandom dot pattern is used to match the fringe center points.Then,the phase information of the points between adjacent fringes is obtained using the Fourier transform.Based on the phase-matching algorithm,to matches all pixels between adjacent matching fringes.Optimizes the matching points to obtain sub-pixel parallax.Finally,the three-dimensional coordinates of the matching points are calculated to obtain the dense three-dimensional reconstruction results of the measured object.(3)The three-dimensional reconstruction technology based on spatially encoded structured light is studied.A structured light 3D reconstruction system was designed and implemented.First,based on the monocular camera calibration method and the binocular system calibration method,the calibration module of the three-dimensional reconstruction system is implemented,which can be used to calibrate the structured light system simply and quickly to obtain the parameters of the structured light system.Then,based on the triangulation principle,the 3D reconstruction module of the system is realized.The structured light pattern proposed in this paper is used to reconstruct the measured object,which can quickly and accurately obtain the 3D data of the object.Finally,a simple denoising process is performed on the 3D point cloud.And realize the visualization of 3D reconstruction results.