学位论文简介

本论文围绕显微关联成像技术的理论构建、系统设计与优化，以及其在活细胞成像与分类中的实际应用，开展了系统性研究，旨在解决传统光学显微成像技术中依赖直接光强探测而导致的局限性，并探索显微关联成像作为无标记、低光损伤成像技术的潜力。取得了以下主要创新性研究成果：

(1) 构建了高分辨率显微关联成像系统模型，并分析其成像性能。理论推导表明，系统分辨率受限于散斑直径，并与离焦距离和波长相关。基于临界分辨率理论，推导了系统的景深公式，发现景深与波长成正比，这与传统成像技术的景深规律显著不同。实验结果验证了系统在微米级物体成像中的高分辨率能力，突破了面阵探测器的像素限制，并显著提高了景深。

(2) 针对活细胞成像的需求，论文进一步优化了系统的硬件与算法。通过引入数字微镜阵列（DMD）作为调制器件，系统的光场调制速率得到大幅提升，同时结合套筒透镜和显微物镜优化了光场缩束效果。论文还设计了匹配 DMD 的信号同步采集方案，结合压缩感知（TVAL3 算法）及 GCS+S 序 Hadamard 矩阵实现了低采样率下的高质量图像重建。此外，提出了基于预处理 S 矩阵的算法，显著增强了抗噪性能并从理论上证明了预处理后S矩阵的行正交性质。

(3) 论文开展了多种活细胞的显微成像实验，成功实现了对人血红细胞、人胚肾细胞（293T）及三种癌细胞（Caov3、Molm13 和 Ishikawa）的高分辨成像和形态学分析。实验验证表明，红细胞的双凹圆盘结构可在30%采样率条件下清晰重建，同时在癌细胞中观察到更高的聚集性与复杂形态差异，为揭示其生物学特性提供了新视角。

(4) 通过基于一维桶探测数据的细胞分类实验，论文展示了显微关联成像在细胞分类识别中的潜力。在极低采样率（0.03%）下即可实现 88% 的分类准确率，并在采样率不超过0.12%时达到了94%的稳定分类性能。这一研究为大规模细胞数据的低成本、高效分析提供了理论依据和技术支持。

主要学术成果

[1] Xiaohui Zhu, Yanfeng Bai, Wei Tan, Liyu Zhou, Xianwei Huang, Tongji Jiang, Teng Jiang, Suqin Nan, and Xiquan Fu. High-resolution microscopic ghost imaging for bioimaging. Physical Review Applied, 2023, 20: 014028.

[2] Xiaohui Zhu, Wei Tan, Xianwei Huang, Xiaoqian Liang, Qi Zhou, Yanfeng Bai, and Xiquan Fu. Noise-robust and data-efficient compressed ghost imaging via the preconditioned S-matrix method. Journal of the Optical Society of America A, 2024, 41(11): 2090-2098.

[3] Xiaohui Zhu, Yanfeng Bai, Wei Tan, Xiaoqian Liang, Qi Zhou, Jian Li, Weijun Zhou, Jintao Zhai, Xianwei Huang, Xiongwei Cai, and Xiquan Fu. Live Cell Imaging and Classification via Microscopic Ghost Imaging. Physical Review Applied (Under review).

[4] 朱孝辉, 梁小茜, 谭威, 黄贤伟, 白艳锋, 傅喜泉. 基于量子关联的显微成像研究进展. 计测技术, 2023, 43(03): 60-74.

[5] Qi Zhou, Yanfeng Bai, Wei Tan, Xiaohui Zhu, Xiaoqian Liang, Jian Li, Xianwei Huang, and Xiquan Fu. Multi-object coaxial imaging with an Airy beam array. Scientific Reports, 2025, 15(1): 4439.

[6] Suqin Nan, Lin Luo, Xuanpengfan Zou, Yang Guo, Xianwei Huang, Wei Tan, Xiaohui Zhu, Teng Jiang, Chuang Li, Yanfeng Bai, and Xiquan Fu. Wide-field scanning ghost imaging based on a local binary pattern and untrained neural network. Optics Express 2024, 32(23): 41644-41656.

[7] Jiaqi Yin, Yanfeng Bai, Liyu Zhou, Xiaohui Zhu, Xuanpengfan Zou, Qi Zhou, Xianwei Huang, and Xiquan Fu. Remote sensing ghost imaging based on Hadamard modulated Gaussian array beam. Optics Communications, 2025, 574: 131108.

[8] Qin Fu, Liyu Zhou, Xianwei Huang, Xiaohui Zhu, Wei Tan, Yanfeng Bai, and Xiquan Fu. Integrated simulation method of the scattering medium. Optics Communications, 2024, 558: 130368.

[9] Xuanpengfan Zou, Xianwei Huang, Wei Tan, Liyu Zhou, Xiaohui Zhu, Qin Fu, Xiaoqian Liang, Suqin Nan, Yanfeng Bai, and Xiquan Fu. Target extraction through strong scattering disturbance using characteristic-enhanced pseudo-thermal ghost imaging. Chinese Optics Letters, 2024, 22(12): 121103.

[10] Liyu Zhou, Yanfeng Bai, Qin Fu, Xiaohui Zhu, Xianwei Huang, Xuanpengfan Zou, and Xiquan Fu. Measurable speckle gradation Hadamard single-pixel imaging. Chinese Optics Letters, 2024, 22(3): 031104.

[11] Wei Tan, Yanfeng Bai, Xianwei Huang, Xiaohui Zhu, Teng Jiang, Xuanpengfan Zou, Suqin Nan, Mingwei Liu, and Xiquan Fu. Feeble-light ghost imaging via correlation calculation. Results in Physics, 2023, 54: 107094.

[12] Teng Jiang, Yanfeng Bai, Wei Tan, Xiaohui Zhu, Xianwei Huang, Suqin Nan, and Xiquan Fu. Ghost imaging lidar system for remote imaging. Optics Express, 2023, 31(9): 15107-15117.

[13] 朱孝辉, 傅喜泉, 白艳锋. 用于血液中异常细胞筛查的基于高速调制随机介质掺杂光纤的关联成像方法. CN202211142180.8. (已授权)

[14] 朱孝辉, 傅喜泉, 白艳锋. 一种基于关联成像的血管中异常细胞不停流成像筛查系统. CN202210043882.4. (已授权)