A Compressive Sensing Based Hyperspectral Ocean Color Imager for CubeSats

Hyperspectral imagery of ocean color from satellites with high spatial (~10 m) and temporal (~hourly) resolution is currently a gap in studying coastal ecosystems. Imaging from CubeSats and high-altitude drones has the potential to fill this gap in the future. These platforms are cost-effective relative to conventional missions with the possibility of broad global coverage through a constellation of such systems.

Hyperspectral imaging is typically carried out with CCD or CMOS array based systems. Some challenges in adapting such imaging technology to CubeSat platforms, especially over coastal regions, include 1) severe data transmission bandwidth limitations, 2) size, weight, power limitations, 3) insufficient signal-to-noise (SNR) for adequate algorithm retrievals, and 4) saturation, blooming and edge effect problems with water adjacent to bright land and clouds.

We are developing a novel pushbroom-type CubeSat imager based on Spatial Light Modulation with a Digital Micromirror Device (DMD). A DMD consists of millions of electrostatic-actuated micro-mirrors that can be used to control light collection dynamically for each individual pixel equivalent. The imager can thus allow for adaptive optimization of aspects such as spectral resolution, spatial resolution, and SNR based on a particular scene being imaged. Front end image compression customized for ocean color optimizes information transmission given severe data bandwidth limits. A single detector such as a sensitive photomultiplier tube or Avalanche photodiode can be used instead of a CCD or CMOS, increasing SNR by potentially orders of magnitude and eliminating intercalibration of detector arrays. The spectral range for the current system covers 340 to 820 nm with a projected spatial resolution of 28 m over a 50 km swath. The imaging system has been prototyped in our lab. Preliminary results from characterization and feasibility demonstrations will be presented.