Research On Integral Imaging Acquisition System Based On Sparse Camera Array

The three-dimensional display technology originated more than a century ago.After more than a century of development,currently,the most widely used three-dimensional display technology in the market is basically implemented using the principle of binocular parallax,which requires viewers to wear corresponding auxiliary equipment.In order to achieve a stereoscopic effect,the viewer may have problems such as inconvenient wearing of the device and visual fatigue.Integrated stereo imaging display technology,as an emerging three-dimensional naked-eye display technology,has the advantages of not causing visual fatigue to viewers,requiring no auxiliary equipment,and using only two eyes to view stereo effects,so it has good development prospects.However,collecting three-dimensional scene information through traditional lens arrays will cause some problems such as inaccurate collection.Therefore,it is of great practical significance to use the combination of camera array acquisition and virtual viewpoint rendering algorithm to build an integrated imaging acquisition system.The use of sparse camera arrays to collect and optimize integrated imaging element images,and improve the generation of virtual viewpoints,can improve the quality of element images obtained using integrated imaging technology,and improve the display effect of the naked eye in three dimensions.In terms of improving the element images obtained by integrated imaging,this paper uses a correction method to improve the element images obtained by the sparse camera array based on the camera position shift problem that may occur in the sparse camera array.Position the camera array before setting the reference point.After calculating the position coordinates of the reference point and the error of the camera position,the relationship between the position error of the camera array and the element image and the accuracy of the correction algorithm are analyzed.Subsequently,the proposed method was validated using optical reconstruction experiments.The results show that this method can effectively reduce the influence of the position deviation of the camera array on the element image array,and make the quality of the reconstructed 3D image better than that before the correction.In terms of rendering virtual viewpoints based on integral imaging technology,this paper proposes a method to optimize the quality of disparity map generation,which can improve the quality of virtual viewpoint synthesis.That is,the image is first color-segmented in the HIS space,and then the modified dynamic programming algorithm is used to perform stereo matching on the image to generate a disparity map.A cost function based on the HIS space is added to the calculation of the matching cost,and the disparity is calculated using the Winner-Takes-All idea to improve the accuracy of the disparity calculation,and then a virtual viewpoint is generated.The experimental simulation results show that compared with the traditional algorithm,the improved algorithm optimizes the quality of disparity map generation,and the quality of the virtual viewpoint image generated by it is also improved,and the information in the image is better retained.