Thursday, January 13, 2011

CS: University of Arizona's Adaptive Spectral Imager

David Brady reminded me on his blog that I made a video explaining the CASSI system:

David also pointed to Michael Gehm's presentation of a similar spectrometer that seems to be performing adaptive compressive sensing while CASSI is non-adaptive in nature.

I like the fact that Mike uses the imagery of group testing to get the message across. Looking back at the blog's archive, I seem to have totally missed it. The corresponding papers are:

We describe the design and development of an adaptive spectrometer that actively manages dynamic range mismatch between source signals and the spectrometer’s detector. The introduction of an inexpensive digital micromirror array yields performance improvements compared to a traditional device. Simulation results are presented that indicate real-world scenarios in which the adaptive spectrometer will yield superior performance. Experimental results from a first-generation prototype adaptive spectrometer confirm the expectations of the simulations.

Spectral imaging is an emerging tool for defense and security applications because it provides compositional information about the objects in a scene. The underlying task—measuring a 3-D dataset using a 2-D detector array—is challenging, and straightforward approaches to the problem can result in severe performance tradeoffs. While a number of ingenious (non-adaptive) solutions have been proposed that minimize these tradeoffs, the complexity of the sensing task suggests that adaptive approaches to spectral imaging are worth considering. As a first step towards this goal, we investigate adaptive spectroscopy and present initial results confirming dramatic cost/performance gains for a particular implementation.

Of particular interest are the next two figures:

I am adding this adaptive spectrometer to the compressive sensing hardware page.

No comments: