Research On Virtual Viewpoint Synthesis Algorithm In Naked Eye 3D Display

Human beings live in a three-dimensional space.Although two dimensional(2D)display technology tends to mature,the traditional 2D planar display technology loses the depth information of the real physical world,which hinders human beings’ perception of the objective world to a certain extent.In order to better meet the requirements of human vision,3d display technology has been widely studied and gradually into our life.However,naked eye 3D display can sense 3D information without wearing any auxiliary equipment,so it needs further research and summary.This paper mainly studies the virtual viewpoint synthesis technology in naked-eye 3D display.The main research is to use discrete viewpoint acquisition and binocular viewpoint acquisition to obtain the original viewpoint image.After obtaining the disparity map,use the virtual viewpoint synthesis technology to generate a new virtual viewpoint.The images were finally reproduced in 3D using an integrated imaging system.The main research contents are as follows:1.The virtual viewpoint synthesis technology based on discrete viewpoint images is studied.In the process of virtual synthesis,the use of depth information has a great influence on the effect of virtual viewpoint generation.Therefore,a depth map acquisition method based on the positional relationship of homologous points is proposed.First,a scanning camera array is used to collect discrete viewpoint images,and depth information is obtained according to the characteristic that the distance between homologous points is equal.Then,a new viewpoint image is generated using the virtual viewpoint synthesis technology based on discrete viewpoint images,and finally the integrated imaging system is used for 3D display.2.The virtual viewpoint content synthesis technology based on stereo matching is studied.To get a higher quality depth image,this paper proposes an adaptive window-guided filtering stereo matching based on Mean shift image segmentation.In the disparity calculation step of the guided filtering stereo matching,the Mean shift image segmentation algorithm is used to obtain the segmented image,and then an adaptive window based on image segmentation is established for the pixels,and the disparity correction is performed on the pixels based on the adaptive window to obtain the disparity map;Then use DIBR technology to obtain a new virtual viewpoint image.