See-Through Light Field Mixed-Reality Display System Based on Polymer Dispersed Liquid Crystal Film
Chen Si,Li Hui,Li Tian,Su Chenbo,Wu Yuntao
DOI: https://doi.org/10.3788/aos240706
2024-01-01
Acta Optica Sinica
Abstract:Objective With rapid development,mixed-reality display systems are increasingly utilized in various fields,such as entertainment,education,and simulation training.However,prolonged use of these systems often leads to visual discomfort for users,posing a significant challenge in this area.The primary cause of user discomfort is the visual accommodation-convergence conflict(VAC),where current mixed-reality display systems often fail to provide the correct focal point.This failure can cause a conflict between the eye's accommodation and convergence mechanisms,leading to visual strain and discomfort.Utilizing liquid crystal elements in see-through,mixed-reality display systems can resolve the VAC problem.However,the fabrication of liquid crystal devices in these methods is complex,and maintaining high precision in manual production processing is also challenging,potentially affecting the quality of the virtual reality.Hence,there is a pressing need for a simple method in enhanced display systems.We propose a novel approach to achieve see-through light field mixed-reality using polymer dispersed liquid crystal(PDLC),specifically designed to address the visual VAC.The proposed system leverages the electrically adjustable transmittance property of PDLC to seamlessly combine the light field data of the projected target in the real world.A light field rendering algorithm is introduced based on convex optimization theory,which effectively fuses the light field redundancy across different PDLC voltages and the color brightness disparities between the actual scene and the target.This integration generates a visually harmonious image containing the virtual and real fused scenes at varying focal depths.Experiments demonstrate that the proposed system achieves superior quality at a continuous focal depth.The proposed system offers several advantages,including reduced computational demands and affordable hardware expenses. Methods The system utilizes PDLC as the core optical element,enabling the combination of light field data from virtual target objects and real scenes.By leveraging the electrically adjustable characteristics of PDLC,the system obtains virtual and real fused light field data at multiple voltages.It implements an adaptive light field rendering algorithm based on convex optimization theory.Two regularization terms are incorporated in the objective function,concerning the device and light field characteristics:A spatiotemporal regularization term considers the redundancy of the captured light field at multiple voltages and uses the redundant information captured at different voltages to refine the details.A harmonized regularization term combines the virtual and real fused light field data captured by PDLC at multiple voltages.It calculates the color brightness difference between the real scene and the virtual object across different color channels,effectively correcting the color of the target virtual object and facilitating seamless fusing with the real scene.The algorithm jointly and optimally reconstructs the multi-voltage virtual and real fused light field data.The data is then decoupled to obtain high-quality two-dimensional images that fuse virtual and real scenes at different depths. Results and Discussions This system investigates the influence of light combiners on see-through mixed-reality display systems,revealing that utilizing PDLC as a light combiner enhances the fusion of reality and virtual elements.This enhancement is attributed to PDLC's compatibility with the proposed light field rendering algorithm(Fig.8).Furthermore,we examine the regularization terms of the light field rendering algorithm,affirming that each term positively influences the outcomes.These terms improve color coordination between the virtual and surrounding real environments,consequently enhancing the overall quality of both virtual and real scene images(Fig.9).To further validate the seamless fusion of real and virtual scenes,we generate coordinated two-dimensional images of real and virtual scenes at various depths of focus(Fig.10).The results of multiple depths of virtual scene fusion,facilitated by reconstruction based on the light field rendering algorithm,show progressive focus on the virtual letters"A","B",and"C"with changes in depth,maintaining consistent quality with the final real scene.Additionally,the proposed system facilitates relatively smooth zoom effects within the real scene.Compared to conventional mixed-reality systems,the proposed system demonstrates significantly superior virtual reality fusion results(Fig.11 and Table 1). Conclusions We introduce a see-through light field mixed-reality system that leverages PDLC to effectively address the VAC issue.Our method utilizes the electrically adjustable properties of PDLC to seamlessly integrate virtual light field information with real scene light field information.Subsequently,a light field rendering algorithm processes multiple sets of fused virtual and real light field data.Using the redundant information at different voltages of the PDLC and the differences between the real and virtual scenes as constraints,the system generates high-resolution,arbitrary focal depth two-dimensional images that blend virtual and real scenes.Extensive experimental validation demonstrates that the proposed system excels in reconstructing the quality of fused virtual and real scene images.It significantly reduces the visual dissonance of virtual images and supports continuous zoom capabilities.This innovative approach offers an enhanced solution to the VAC problem in existing see-through mixed-reality displays.