随着现代科技对纳米微观区域兴趣的增加, 如DNA测序、 分子纳米器件微结构检测等, 其对拉曼光谱技术的空间分辨力提出了更高的要求, 而现有共焦拉曼光谱技术受自身原理限制, 空间分辨力已无法满足科学需求。 针对这一问题, 在现有共焦拉曼光谱技术的基础上, 提出一种基于最大似然算法的共焦拉曼光谱成像方法。 该方法将超分辨图像复原技术与共焦拉曼光谱技术相结合, 利用基于Poisson-Markov约束的最大似然超分辨复原算法对共焦拉曼光谱图像进行超分辨图像复原处理, 恢复图像高频成分, 进而改善共焦拉曼光谱系统的空间分辨能力, 实现超分辨成像。 仿真分析和实验结果表明, 提出的基于最大似然算法的共焦拉曼光谱成像方法在不改变现有共焦拉曼光谱系统光学结构的前提下, 仅对单幅拉曼光谱图像进行超分辨图像复原处理, 即可将系统空间分辨力提高到200 nm, 实现超分辨成像, 同时该方法具有较强的噪声抑制能力。 该方法有效地提高了共焦拉曼光谱系统的空间分辨力, 为物理化学、 材料科学等前沿领域中的高空间分辨微区光谱探测提供了一种新的途径, 是一种行之有效的高空间分辨的共焦拉曼光谱成像方法。

With the increasing interest in nano microscopic area, such as DNA sequencing, micro structure detection of molecular nano devices, a higher requirement for the spatial resolution of Raman spectroscopy is demanded. However, because of the weak Raman signal, the pinhole size of confocal Raman microscopy is usually a few hundreds microns to ensure a relatively higher spectrum throughput, but the large pinhole size limits the improvements of spatial resolution of confoal Raman spectroscopy. As a result, the convential confocal Raman spectroscopy has been unable to meet the needs of science development. Therefore, a confocal Raman image method with Maximum Likelihood image restoration algorithm based on the convential confocal Raman microscope is propose. This method combines super-resolution image restoration technology and confocal Raman microscopy to realize super-resolution imaging, by using Maximum Likelihood image restoration algorithm based on Poisson-Markov model to conduct image restoration processing on the Raman image, and the high frequency information of the image is recovered, and then the spatial resolution of Raman image is improved and the super-resolution image is realized. Simulation analyses and experimental results indicate that the proposed confocal Raman image method with Maximum Likelihood image restoration algorithm can improve the spatial resolution to 200 nm without losing any Raman spectral signal under the same condition with convential confocal Raman microscopy, moreover it has strong noise suppression capability. In conclusion, the method can provide a new approach for material science, life sciences, biomedicine and other frontiers areas. This method is an effective confocal Raman image method with high spatial resolution.