Research On Image Generation Technology In 3D Display

Humans use eyes capture the depth information of the objects in the real world and then build three-dimensional effect in mind.The images on traditional two-dimensional display are short of depth information,so three-dimensional displays which can restore the depth information of images and the three-dimensional image generation technology of virtual scene have received the widespread attention of researchers both at home and abroad.At present,there are still shortcomings in the image generation technology suitable for three-dimensional display devices.In this paper,the three-dimensional image generation technology of virtual reality head-mounted display and autostereoscopic display will be studied.The main research contents and innovations of this paper are as follows:1)Discuss the principle of image generation of virtual reality head-mounted display,analyze the reason why the near-eye lens of virtual reality head-mounted display will cause the observation image to show pillow distortion,and propose a generation method of barrel distortion image based on geometric vertices.After the derivation of the vertex distortion formula,the experimental program is written by using C++programming language and GLSL shader language,and the distortion effect verification experiment is carried out.This method realizes image distortion and rendering through vertex coordinate transformation in a single rendering pipeline process.Compared with the post-processing method,the barrel distorted image obtained by this method has better geometric continuity and more accurate pixel color.2)Discuss the principle of image generation of autostereoscopic display,analyze the phenomenon of monotonous lack of three-dimensional display content caused by the limitation of depth of field and computer hardware,and propose a method of generating multi-projection and multi-view images.Based on the multi-projection composition method,two kinds of multi-projection virtual scenes are constructed on Unity3D graphics rendering engine,the corresponding multi-projection multi-view images are generated,and the verification experiment of display effect are carried out on the autostereoscopic display.Compared with the existing 3D display content of autostereoscopic display,this design broadens the display application field of autostereoscopic display,and can present three-dimensional display effect for multi-projection virtual scene.The research results of this paper provide a new solution to the problem of pillow distortion in the image rendering process of virtual reality head-mounted display,which has practical engineering significance.It also enriches the three-dimensional display content of the autostereoscopic display,so that it can render multi-projection virtual scenes,which lays a good foundation for the production of three-dimensional animation suitable for the autostereoscopic display.