计算多光谱视频成像[特邀报告]
Computational multispectral video imaging [Invited].
作者信息
Wang Peng, Menon Rajesh
出版信息
J Opt Soc Am A Opt Image Sci Vis. 2018 Jan 1;35(1):189-199. doi: 10.1364/JOSAA.35.000189.
Multispectral imagers reveal information unperceivable to humans and conventional cameras. Here, we demonstrate a compact single-shot multispectral video-imaging camera by placing a micro-structured diffractive filter in close proximity to the image sensor. The diffractive filter converts spectral information to a spatial code on the sensor pixels. Following a calibration step, this code can be inverted via regularization-based linear algebra to compute the multispectral image. We experimentally demonstrated spectral resolution of 9.6 nm within the visible band (430-718 nm). We further show that the spatial resolution is enhanced by over 30% compared with the case without the diffractive filter. We also demonstrate Vis-IR imaging with the same sensor. Because no absorptive color filters are utilized, sensitivity is preserved as well. Finally, the diffractive filters can be easily manufactured using optical lithography and replication techniques.
多光谱成像仪能够揭示人类和传统相机无法感知的信息。在此,我们展示了一种紧凑的单次多光谱视频成像相机,通过将微结构衍射滤光片放置在靠近图像传感器的位置。衍射滤光片将光谱信息转换为传感器像素上的空间编码。经过校准步骤后,该编码可通过基于正则化的线性代数进行反演,以计算多光谱图像。我们通过实验证明在可见光波段(430 - 718纳米)内光谱分辨率为9.6纳米。我们还表明,与没有衍射滤光片的情况相比,空间分辨率提高了30%以上。我们还用同一传感器展示了可见 - 红外成像。由于未使用吸收性滤色片,灵敏度也得以保留。最后,衍射滤光片可使用光刻和复制技术轻松制造。