Research On Glasses-free 3D Image Generation Algorithm And Technology Based On Binocular Vision

Glasses-free 3D is the frontier and future development trend of display technology.However,at present,the production of high-quality 3D content is complex and costly,which hinder the development of glasses-free 3D.Aiming at this problem,based on binocular stereo vision technology,this paper makes stereo rectification for binocular image obtained by dual cameras,and uses binocular matching algorithm to obtain disparity map,so as to generate 3D image.The main work is as follows:1.Obtain color + gray binocular image,and study the binocular camera calibration and stereo rectification technology.Based on Zhang's camera calibration method and Bougouet method,complete the calibration of the dual cameras of the mobile phone used in the experiment and the stereo rectification of the binocular image by using Matlab and OpenCV.2.Research on binocular matching algorithm.Aiming at the situation that light distortion is easy to occur in the actual scene shooting and mismatching is easy to occur in the region with low texture or disparity discontinuity,a matching cost calculation method of gradient and Census transform adaptive weighted fusion and a cost aggregation algorithm of combining guiding filter and adaptive window are proposed respectively.The proposed algorithm was tested and quantitatively analyzed using Middlebury's standard binocular image.3.To further verify the applicability of the proposed matching algorithm to the real scene,the proposed algorithm is applied to the binocular image of real scene to generate a disparity map.The color left image and the disparity map are used to generate a color right image by forward mapping and backward mapping,and the color binocular image is synthesized into a 3D image.In this paper,a dual-camera 3D mobile phone,ZTE Tianji 7MAX,is selected as the hardware platform,and the simulation experiment of the proposed algorithm is completed on the VS2017 platform.The experimental results show that the 3D images generated by the algorithm proposed in this paper have better performance in viewing comfort and stereoscopic display than the 3D images taken by the phone.The algorithm proposed in this paper can reduce the complexity and cost of the production of high-quality 3D resources and has certain practical application value.