集成式仿生复眼光引擎驱动的三维成像微投影显示系统研究

At present, micro-displays develop towards to the miniaturization, visualization and intelligentization. However, the problem such as low light-energy utilization and the lack of the ability of 3D imaging becomes a bottleneck for further development of micro-projection displays. According to the urgent needs of the miniaturization and intelligentization for micro-projection displays, this project intends to develop a novel 3D bioinspired pico-projection driven by the light engine with artificial compound eyes. First, a multi-factor coupling model with cross-scale will be established for the light engine on the basis of the theory of coupling bionics. Second, freeform lens will be studied with the theory of simultaneous multiple surfaces (SMS). Third, the structure of the freeform light engine with artificial compound eyes will be developed in accordance with the relationship between the light source and the target surface in the topological space. At the same time, the adaptive network-based fuzzy inference system (ANFIS) can adjust and optimize the structure parameters by using expert system, something that it enhances the control precision and the function of self-adaptive and self-learning. Moreover, a femtosecond-laser-induced multi-beam parallel process and a micro-embossing process will be developed in the project, in order to fabricate large areas, uniform nano-structures on flat substrates and achieve the integration of the light engine. Finally, a 3D bioinspired pico-projection will be established and applied to dynamic display and interactive virtual reality. As a multi-discipline research on manufacturing engineering and bio-nanotechnology, this project will provide theoretical and technical supports to novel micro/nano bionic devices in the future.

随着微型投影显示系统便携化、智能化和立体视觉化的发展，其光能利用率低及立体图像还原能力的不足等问题日益凸显。本项目拟通过借鉴蜜蜂复眼器官微纳结构特征及其视觉感知信息传导机理，研发一种集成式仿生复眼光引擎驱动的三维微投影显示系统。基于耦合仿生原理，建立三维复眼光引擎器件跨尺度多元耦合模型；采用多面同步方法设计加载复眼器件的自由曲面基底；依据光源出射光线矢量方向与目标面的拓扑映射关系，通过自适应神经模糊算法进行面型参数自学习和自调整，集成基于自由曲面的复眼光引擎一体化结构；开展基于飞秒脉冲多光束并行加工的复眼光引擎加工和微压印工艺研究，探索仿生复眼微纳器件批量化工艺路线；最终研制出实现实时动态的传输显示信息和交互式的生境再现功能的三维立体成像的微投影显示系统。本项目旨在通过制造科学与生物技术的交叉研究，为进一步发展仿生微纳器件奠定理论和应用基础。

本项目通过借鉴蜜蜂复眼器官微纳结构特征及其视觉感知信息传导机理，研发一种集成式仿生复眼光引擎驱动的空间光场相机成像系统。基于耦合仿生原理，建立三维复眼光引擎器件跨尺度多元耦合模型，通过自适应神经模糊算法进行面型参数自学习和自调整，集成基于自由曲面的复眼光引擎一体化结构。开展基于飞秒脉冲多光束并行加工的复眼光引擎加工和微压印工艺研究，复眼单颗小眼直径为79.160微米，表面粗糙度10纳米，尺寸精度为0.011微米。深入开展复眼器件与光场相机耦合设计，最终研制出实现实时动态的传输显示面阵微镜阵列成像的微透镜阵列光场相机显示系统。该微透镜阵列型光场相机结构特点包括，在主镜头与图像传感器平面中放入复眼结构微透镜阵列，每个微透镜阵列下在传感器的成像都对应不同视角下的场景信息。光场相机可在三维模式下提取图像的形貌特征信息，进而深度学习能够在形貌模态数据基础上通过卷积主动学习特征。利用仿生微透镜阵列光场相机显示系统，聚焦大气颗粒物散射光强度、偏振信息及微观形貌多模态特征信息有效提取的测量技术，围绕提取大气颗粒物散射光强度、偏振信息及微观形貌的多模态信息，分析了各模态下大气颗粒物探测数据的有效表达，运用图像变化思想，对各种典型大气颗粒物各向图像变化结果进行估计而获取基于形貌分形学的形貌特征矢量，进行特征划分与分类后完成形貌提取。光场相机获取原始图像，经过滤波去噪、阈值处理及梯度修正后完成图像预处理后，计算源图像的形貌分形值，选取特征区域值得到特征矢量。综合完成了对三维散射场频域图像采集，根据分形维及数学形态维的估计方法提取颗粒形貌特征值参量模态。解决了颗粒物测量仿生器件多传感器融合的材料-结构-功能一体化设计制造问题，提高测量器件全散射角的测量范围和识别精度，实现多模态特征信息的有效提取。本项目通过仪器制造科学与生物技术的交叉研究，丰富了仿生复眼器件设计和加工理论，为进一步发展仿生微纳探测器件奠定了理论和应用基础。