全息术搭建了光学成像和光散射领域的桥梁，作为一种无标记成像技术，为物理学、生物学和材料科学领域提供了有效的观测维度，是定量测量复杂光场分布的重要工具。当今迅速增长的即时检验、远程医疗等应用，需要成像系统具备无标记、多维度、高分辨、宽视场等多方面的性能提升。本文围绕高保真、高通量、无标记、三维重建等成像需求，系统研究了高空间带宽数字全息光学成像技术。提出了一种基于双平面耦合的高空间带宽全息成像方法，基于交替投影的相位恢复框架，证明了离轴优化相位可提供有效的初始解，能避免迭代计算停滞问题以及收敛性问题。结合同轴全息和离轴全息，保持了高空间带宽、高保真度的优点，实现了基于双衍射平面的快速定量复振幅重建。计算过程中无需对样本施加先验假设，包括支撑约束、非负约束、稀疏约束等，可实现定量复振幅显微成像。提出了一种基于单边带解析方法的高空间带宽全息复用解耦框架，针对全息图空间带宽利用率受限的问题，构造了复用函数的解析条件，使多物光在复用时无需与全息图的零级空间频谱分离，进一步利用了全息图冗余的空间频谱，将单帧全息复用的空间带宽利用率从58.9%提升至78.5%。基于该框架设计了高空间带宽数字全息显微系统，在不改变数据采集量和空间分辨率的前提下，达到了常规全息显微镜8倍的成像空间带宽积。基于交替方向优化方法提出像差自适应优化理论，将像差提取作为凸优化问题，对像差以泽尼克多项式为基底进行直接拟合与分离，实现了高空间带宽且低像差的定量相位成像。提出了高空间带宽三维层析成像与细胞追踪方法，构建了双光束复振幅解析函数，将傅里叶衍射理论与复用传递函数结合，规避了衍射层析中由于单张全息图空间带宽受限所造成的成像空间通量限制。实验中，对无标记的肠内分泌肿瘤的各种类型进行了三维重建与分析，并将全息成像应用至微流控单细胞分析中，提出了三维时空编码粒子追踪方法，规避了高速成像中的计算冗余，实现了时空单细胞运动轨迹追踪，降低了跟踪时间。
Holography provides a possibility to bridge the gap between the optical imaging and light scattering. It is a label-free imaging method, providing an effective observation dimension for the fields of physics, biology, and materials science. It has become a powerful tool for optical complex wavefront measurement. Nowadays, rapidly growing demand for applications in instant inspection and telemedicine requires imaging systems to have performance improvement such as label-free, multi-dimensional, high resolution and wide field of view. Focusing on the imaging requirements of high fidelity, high throughput, label-free, and three-dimensional reconstruction, the specific research work is as follows:A high space-bandwidth holographic imaging based on the non-prior coupled phase retrieval is proposed. Off-axis optimization phase provides an effective initial guess to avoid the iterative computation stagnation and low convergence based on alternating projection phase retrieval framework. By combining inline holography and off-axis holography, it can maintain the performance of high space-bandwidth and high fidelity, and realize rapid quantitative complex-amplitude reconstruction based on two diffraction planes. The proposed strategy works well without any prior assumptions of the objects including support, non-negative, sparse constraints, etc. Quantitative complex amplitude microscopic imaging can be realized.A general decoupling framework for high space-bandwidth holographic multiplexing based on single sideband analysis method is proposed. To break the limited space bandwidth utilization of hologram, an analytic condition of the multiplexed wavefront is constructed. The spatial spectra of multiplexed wavefronts do not need to be separated from the zero-order spectrum of the hologram. The bandwidth utilization of single multiplexed hologram in a diffraction-limited optical system can reach from 58.9% to 78.5%. Based on this framework, a digital holographic microscope system with high space-bandwidth is designed. It enables 8-fold space bandwidth product (SBP) enhancement of the reconstruction without sacrificing the frame rate and spatial resolution. An adaptive optimization framework of quantitative phase aberration extraction based on alternating direction method is proposed. By formulating the aberration extraction as a convex quadratic problem, the background phase aberration can be fastly and directly decomposed with the complete basis functions of Zernike polynomials. A high-bandwidth and low-aberration quantitative phase imaging can be achieved.The high space-bandwidth 3D optical diffraction tomography and 3D cell tracking method were proposed. Based on the analyticity of multiplexed complex-amplitudes, the relation between Fourier diffraction theory and multiplexing transfer function is established. The throughput restriction of the space-bandwidth of single hologram in interferometric-based diffraction tomography can be circumvented by the proposed framework. Different 3D tumor types and a variety of precursor pathologies can be visualized with label-free. By combining holographic imaging to single-cell microfluidic analysis, a 3D spatiotemporal coding tracking method is proposed. The computational redundancy in high-speed imaging can be circumvented. The spatiotemporal single-cell tracking in microfluidic is realized. The tracking time can be reduced.
