Beating Nyquist: Compressed Sensing in Astronomical Instrument Design
 

The conventional approach in astronomical instruments is to sample the signal at more than twice the Nyquist frequency; for example, the pixels in an imager are generally less than half the full-width half maximum of the point spread function, and in spectroscopy spectral pixels are less than half the full-width half maximum of the line spread function. This leads to a requirement for very large numbers of pixels in order to sample a wide-field image or a broad-band spectrum. However, recent developments in data science have shown that for certain types of signal, it is possible to significantly under-sample measurements and recover the full signal at high fidelity using a technique called compressed sensing. 

In this talk I will describe this technique, and its applicability to astronomy. I will discuss potential applications to the design of astronomical instruments with optimised sampling, including imaging, spectroscopy, and wavefront sensing, with far fewer pixels than conventional instruments.