将分子成像技术和高光谱技术相结合, 研制了基于AOTF(Acousto-optic Tunable Filters)的分子高光谱成像系统。 系统由显微镜、 分光仪、 CCD镜头、 图像数据采集卡和计算机等几部分组成。 在综合考虑各功能部件的性能及相互的制约关系的基础上, 分析了系统的性能指标, 系统的光谱范围从550～1 000 nm, 可采集200个波段, 空间分辨率可达0.061 5 μm, 光谱分辨率可达2 nm, 当CCD工作在积分模式下采集速度可达到2.612 5 s·B-1, 当CCD工作在非积分模式下可达到约0.11 μs·B-1。 由于受系统光源和光路中透镜及传感器性能的影响, 采集到的图像数据需要进行预处理, 文中提出一种空间维和光谱维联合校正的灰度校正系数算法, 并给出算法的具体实现。 以白血病的血液作样本, 通过对比校正前后的单波段图像、 伪彩色图像和光谱曲线, 说明校正算法的有效性, 为后续的光谱图像数据分析提供了有效的数据。

Integrating molecular imaging technology and hyperspectral technology, a novel molecular hyperspectral imaging (MHSI) system based on AOTF was presented. The system consists of microscope, spectrometer, matrix CCD, image collection card and computer. The system’s performance was synthetically evaluated referring every part’s performance. The spectral range of the MHSI system is from 550 to 1 000 nm. Two hundred twenty five bands can be continuously captured at a time. The spectral resolution is less than 2 nm. The spatial resolution is about 0.061 5 μm. CCD acquisition speed achieved 2.612 5 s·B-1 in the integration mode and about 0.11 μs·B-1 in the non-integration mode. Due to the infection of lamp, a spectral curve extracted directly from the original hyperspectral data can not truly present biochemical character and needs to be corrected. The paper proposes the gray correction coefficient algorithm with spatial dimension and spectral dimension, and gives concrete realization of the algorithm. Taking the sample of leukemia blood, by comparing the single-band images, pseudo-color images and spectra before and after correction, the results indicate the effectiveness of correction algorithm. The corrected data is effective for subsequent analysis.