Research On Three-dimensional Reconstruction Algorithm With High Accuracy Based On Computational Integral Imaging System

With the development of science and technology,three dimensional information processing applications have been more and more popular,for example 3D movies,3D printing and 3D medical imaging.Computational Integral Imaging is well known by its advantages like full-parallax,continuous viewpoint,no need for coherent light source and auxiliary equipment.In the process of three dimensional information representation,computational reconstruction algorithm with high precision is crucial.The algorithms will influence the reconstruction results directly when under the same recording system parameters.In this paper,the reconstruction process of threedimensional profile representation and depth image acquisition has been studied.Based on this,a statistical 3D points reconstruction method with high precision was proposed to remedy the reconstruction accuracy influenced by corresponding points' miss-match.At the meantime,depth image reconstruction method based on axially moving Integral Imaging was proposed which can reconstruct the 3D scene with high image quality.The main contents of this thesis are summarized as follow:Firstly,according to the principle and computational procedure of threedimensional points profile representation based on Integral imaging system,the influence of corresponding points' miss match was analyzed.On account of the large number of elemental images recording by Integral Imaging,statistical method was utilized in the depth image reconstruction process to improve the image quality.Secondly,simulation and optical experiments were implemented to verify the effectiveness of the proposed statistical 3D points reconstruction method.The result shows that it can accurately calculate the 3D coordinates of the object and the 3D profile representation was obtained.Thirdly,the advantages of the axially moving Integral Imaging system was analyzed which is easy to realize and obtain more information.The two traditional reconstruction methods named as superposition reconstruction and pixel mapping with interpolation have its own drawbacks.Therefore,they are not appropriate to the axially moving system.The pixel mapping with back projection reconstruction method was proposed which can reconstruct the depth image with high accuracy.Simulation optical experiments was carried out and the depth image of the scene with high accuracy was obtained.