Research On Full-chain Characterization Of 3D Integral Imaging Display Quality And Light Field Acquisition/Conversion Technology

Integral imaging display,as a kind of "true" three-dimensional display technology,can break through the defects of the traditional 2D display and reconstruct 3D object images.When the resolution,depth,viewing angle,and dynamic refresh rate and other technical indicators of 3D display reach a certain level,it can make the viewer seem to be immersed in the real space,and has broad application prospects in the military and civilian fields.However,there is still a big gap between the current level of integral imaging 3D display and the actual application requirements.In order to realize the development of high-quality integral imaging 3D display system,this paper focuses on the comprehensive characterization of the full-chain integral imaging and large-scale integral imaging video acquisition and conversion technologies.For integral imaging 3D display systems based on symmetrical micro lens arrays,sparse camera arrays and light field cameras,high-quality light field image acquisition and conversion generation methods under different light-field acquisition modes were studied.The specific research work carried out is as follows:(1)Stereoscopic perception characteristics and the method of integral imaging performance characterization were studied.The human visual system is the receiver of 3D display images and introduced from two aspects: human eye optical model and stereoscopic visual characteristics.It simulates Arizona human eye model and determines the requirements for 3D display resolution.A comprehensive overview of the integral imaging 3D display quality evaluation system composed of 3D voxel,system observable depth,and field of view is presented,which provides a theoretical basis for the establishment of a full-chain integral imaging 3D display quality comprehensive characterization model.(2)The principle of optical information transfer in the full links(acquisition process,light field conversion and display process)for integral imaging system is studied.Based on the theory of wave optics,a 3D point-diffusion model of integral imaging system is deduced,and the light intensity distribution characteristics of the 3D reconstruction image are analyzed.Based on the theory of geometrical optics,considering the effects of micro-lens diffraction,imaging distortion,detector sampling effect,display pixel discretization effect,display parameter mismatch,and human visual characteristics,a full-chain integral imaging 3D display quality characterization model was established.The influence of the key system parameters on the 3D display quality is discussed.The full-chain model can optimize the design of the integral imaging system.(3)The inferior resolution of 3D imaging is one important limitation for the development of integral imaging(II).This paper proposes two novel super-resolution 3D II methods using the time-multiplexed coding mask or randomly coded aperture for improving the resolution of 3D imaging based on compressive sensing(CS)theory.Instead of sensing raw pixel data,the recording device measures the compressive samples of the observed 3D scene through the coding mask in 3D II system.The super-resolution EIs array is further recovered by the optimization algorithm.These methods break through the limitations of the traditional integral imaging acquisition mode and the resolution of detector array,which effectively improve the 3D display quality.(4)In order to achieve large-scale light field data acquisition and conversion generation of real three-dimensional scenes,this paper designs and builds a sparse camera array integral imaging acquisition system.Four researches have been carried out from the fields of sparse camera array calibration and correction,depth acquisition based on binocular disparity,generation of virtual view elemental images,and conversion of integrated display light field parameters to match display parameters.An elemental image array generation method is proposed,which combines virtual view image array generation method with a preset depth and the preset viewpoint ray tracing method.The research results can generate the elemental image array that matches any display device parameters.(5)For the light field camera acquisition system has the advantages of lightweight and portable,without calibration,we designed and developed a focused light field camera prototype and capture the light field image.The 3D point cloud image reconstruction algorithm based on depth assessment and the high-resolution auto-focus image reconstruction algorithms are used to reconstruct the collected light field images.The image computational reconstruction algorithm can avoid degradation of the image quality caused by aberrations and assembly errors of the display device.The reconstruction result can be used as an objective evaluation criterion for the 3D display quality.