Spectral Imaging

Instead of typical digital cameras that have sensors that divide an image into 3 bands that mimic humans’ trichromatic vision, Spectral Imaging will increase the number of bands so more information is gleaned about the underlying spectral information of the light. Images can extend into the non-visible parts of the spectrum.

  • Multispectral Imaging typically uses 4~15 spectrum bands and is achieved with a standard camera with RGB sensor paired with different kinds of illuminants and filters.
  • Hyperspectral Imaging uses a dedicated Imaging Spectrometer and is capable of many more bands, typically 64-512, and expands.

When predicting the effect of filters (e.g. EnChroma glasses) or illuminants (see CRI) on trichromatic vision, knowing the complete spectral content is necessary. They are also necessary to simulating how different visual systems (e.g. of different animals) would perceive a given scene.

A hyperspectral image comprises several sub-images (left), each band-passed to a specific wavelength. Each pixel of the image therefore comprises its own spectrum (right)