Research On Computational Reconstruction Technology Of 3D Integral Imaging

With the continuous development of society and the continuous improvement of people's living standards,traditional two-dimensional planar image information can no longer meet people's life needs.Three-dimensional scene information has attracted increasing attention and research.Structural information acquisition and other aspects have great development potential and application prospects.Three-dimensional integrated imaging belongs to the field of naked-eye true three-dimensional display.Although it is an emerging stereoscopic display technology,with its own advantages,such as full parallax,full color,continuous viewpoint display,and naked eye viewing without visual fatigue,it attracts More and more domestic and foreign research scholars have carried out in-depth research and exploration on it,and it has gradually become the key research direction of stereo display technology.Three-dimensional integrated imaging computational reconstruction is to use a computer to simulate a microlens array or a pinhole array,so that it can obtain effective threedimensional scene information from the microunit image array through a related reconstruction algorithm,so that the reconstruction can be calculated 3D scene image.Threedimensional integrated imaging computational reconstruction mainly studies two aspects: free viewpoint computational reconstruction and three-dimensional depth surface computational reconstruction.Due to the current low resolution problem of free viewpoint calculation and reconstruction,on the basis of in-depth research on it,this paper proposes a new method of integrated imaging calculation and reconstruction based on non-redundant image block extraction,which determines the position of the viewpoint In the case of highresolution free-view calculation reconstruction images can be generated.First,according to the position of the viewpoint,determine the position and size of the image block extracted from each micro-unit image;second,calculate the pixel coordinates and the number of pixels contained in the image block extracted from each micro-unit image,and extract each image block The included pixels are rotated 180 degrees and sequentially stitched to reconstruct a high-resolution free-view calculation reconstructed image.At the same time,due to the current low accuracy of 3D depth surface calculation and reconstruction,on the basis of in-depth research on it,this paper proposes a depth surface reconstruction algorithm based on pixel rendering-microlens imaging model,which is effective It avoids the problems of the traditional depth surface calculation and reconstruction algorithm itself,so that the accuracy of the depth surface reconstruction image of the 3D scene can be improved.