A single-shot hyperspectral camera has been built and a single-shot ultrafast videography protocol has been developed. The camera is capable of retrieving three dimensions of information (two spatial dimensions along with a time or spectral dimension) from a two-dimensional signal of a conventional CCD camera. The third dimension is retrieved using a combination of compressed sensing along with the spatial and spectral modulation of an incoming signal.
Time resolution using the camera is achieved by passing a chirped probe pulse through a temporally interesting event. The compressed sensing algorithm then reconstructs a hyperspectral datacube, where slices of the datacube that are lower frequency (redder) will correspond to an earlier point in time while higher frequency slices (more blue) correspond to a later point in time. We expect to capture a video at tens of trillions frames per second.