透过散射介质的宽谱照明非侵入式成像机理研究

Realizing noninvasive real-time imaging through scattering media has great physical significance and application value for many areas, like biomedical imaging, public security and anti-terrorism. Light will be scattered when it propagates through the inhomogeneous media, such as tissues, cloud, and so on, which determines that we cannot observe the expected targets through such materials with conventional optical imaging system. Fortunately, a latest proposed method in 2014, called speckle correlation imaging, has a great performance on observing the hidden targets behind scattering media. However, this method has some obvious drawbacks: first, it needs the narrow band light as the illumination source, which limits its applications; second, the image reconstruction algorithm in the presented method has low convergence rate and poor robustness, which will affect the imaging efficiency. In regard to this, based on the optical memory effect and linear superposition of the point spread function (PSF) of the linear imaging system, this project proposes a new method to realize real-time noninvasive imaging through scattering media with broadband light source. Further, in order to ensure the convergence rate and speed up the imaging process, the designed image reconstruction method in this project employs the sparse information as the prior knowledge and then combines with the compressed sensing theory to reduce the sampling data. Finally, realize the real-time monitoring and imaging for moving targets by studying on the correlation of time-series speckle patterns. This project can thoroughly illuminate the principles of noninvasive imaging through scattering media with broadband light source, which will extend the applications of scattering imaging further.

透过散射介质实现非侵入式实时成像在生物医学成像、公共安全和军事反恐等领域中具有重要的物理意义和应用价值。光波在透过生物组织、云雾等介质时会产生散射现象，导致传统成像系统无法透过此类散射介质对感兴趣的目标清晰成像。2014年最新提出的散斑相关成像方法能够很好地透过散射介质观测到隐藏目标，但该方法过于依赖窄谱光源进行照明，重建算法效率低、鲁棒性差。本申请提出以光学记忆效应为基础，依据线性成像系统点扩散函数的线性叠加关系，构建宽谱光源分段成像的非侵入式散射成像模型；以目标稀疏度作为重建先验信息，结合压缩感知理论设计降采样图像重建算法，提高重建算法收敛率和成像速度。最后，可以通过研究序列散斑图像之间的相关特性，实现对运动目标的实时监测与追踪。该研究能够深入揭示宽谱光源照明下透过散射介质的非侵入式成像机理，拓宽散射成像的应用领域。

透过散射介质实现非侵入式实时成像在生物医学成像、公共安全和军事反恐等领域中具有重要的物理意义和应用价值。本申请以光学记忆效应为基础，依据线性成像系统点扩散函数的线性叠加关系，构建了宽谱光源分段成像的非侵入式散射成像模型；以目标稀疏度作为重建先验信息，结合压缩感知理论设计降采样图像重建算法，提高了重建算法收敛率和成像速度。最后，通过研究序列散斑图像之间的相关特性，实现了对运动目标的实时监测与追踪。该研究能够深入揭示宽谱光源照明下透过散射介质的非侵入式成像机理，拓宽散射成像的应用领域。