High Dynamic Range Surface Object 3D Reconstruction Technology Based On Structured Ligh

Three-dimensional reconstruction technology can obtain the three-dimensional model of the object to be measured by getting point cloud data with three-dimensional measurement technique and processing the point cloud data.It has a wide range of applications in various fields of real life,such as industrial detection,reverse engineering,cultural relic protection and clinical medicine.The three-dimensional reconstruction technology based on structured light is commonly used due to the advantages of high precision and fast speed.However,due to the wide range of surface reflectance of the high dynamic range surface object,the traditional three-dimensional reconstruction technology cannot reconstruct the object with high dynamic range surface effectively.In order to solve the problem of three-dimensional reconstruction of high dynamic range surface object,the high dynamic range surface object three-dimensional reconstruction technology based on structured light is proposed in this paper.The technology regulates the initial exposure time of the camera quickly according to the projection intensity ratio when the object image does not contain saturated pixel,so as to obtain the ideal gray distribution data.Gaussian fitting was carried out to calculate the optimal exposure time of different reflectance regions quickly and accurately.The twelve-step phase shifting fringe patterns are projected under the optimal exposure time to reduce the nonlinear error of measurement results.The point clouds registration algorithm based on key points feature description and the multi-view point clouds registration algorithm based on grouping iterative optimization are used to reduce the registration error and realize high-precision three-dimensional reconstruction.In order to verify the effectiveness of the high dynamic range surface object threedimensional reconstruction technology based on structured light,several three-dimensional reconstruction experiments of high dynamic range surface objects are carried out in this paper.Compared with the traditional phase shifting profilometry,the proposed technology increases the effective point cloud sets of the object to be measured by 50.35% on average and improves the measurement accuracy by reducing the nonlinear error caused by Gamma effect.Moreover,this paper improves the traditional point clouds registration algorithm and improves the accuracy of point clouds registration.The experimental results show that the proposed technology can effectively complete three-dimensional reconstruction of high dynamic range surface objects.