Jonathan Oliver, Ted Roush, Paul Gazis, Wray Buntine, Rohan Baxter, Steve Waterhouse
In the near future NASA intends to explore various regions of our solar system using robotic devices such as rovers, spacecraft, airplanes, and/or balloons. Such platforms will carry imaging devices, and a variety of analytical instruments intended to evaluate the chemical and mineralogical nature of the environment(s) that they encounter. The imaging and/or spectroscopic devices will acquire tremendous volumes of data. The communication bandwidths are restrictive enough so that only a small portion of these data can actually be sent to Earth. The aim of this research was to develop a system which analyses rock spectra to automatically determine which spectra are interesting, and to compress the spectral data for communication to Earth. In the research we report here we classify laboratory data using clustering techniques (ACPro, an enhanced version of Autoclass) and provide the planetary scientists with a rapid, visually oriented method of evaluating the underlying chemical and mineralogical information contained within the clusters. We show how clustering can be used to identify interesting rock samples and estimate the compression that using such a system can achieve.