Spatial light modulation and optical mixing for snapshot hyperspectral imaging architectures

Hyperspectral imaging (HSI) systems measure and partition three dimensions of incident light to construct a hyperspectral image-constituting of two spatial dimensions and one spectral dimension. Rapid acquisition of hyperspectral images is challenging as an image sensor can only measure two dimensions of incident light directly, leaving the remaining dimension of incident light to be acquired over time via scanning processes. This challenge is addressed by the development of snapshot HSI architectures that rapidly and concurrently acquire all three incident light dimensions. In this work, a snapshot HSI architecture is presented which forgoes time-based scanning processes through use of a single-single-point photodiode sensor and liquid crystal (LC) pixel shutter arrays for amplitude modulation. Here, incident light is partitioned into discrete image pixels and frequency encoded by projection onto spatial and spectral modulation LC pixel shutter arrays, with resultant sum and difference frequency components manifesting from optical mixing. The single-point sensor diode transduces the optically mixed light into an electric signal for post processing. A hyperspectral image is reconstructed by means of a Fourier-based frequency analyses, which uncovers the sum and difference frequency components according to the associated LC pixel shutter modulation frequency. A filter bank is used to isolate each spatial and spectral pixel modulation frequency from the Fourier-domain before the associated remapping into the hypercube, being a tensor data structure. The presented snapshot HSI architecture is investigated in terms of the underlying theoretical framework, optical system geometry, with a demonstration of operation through computer simulation and laboratory experimentation.