Project Goal

To develop a general purpose statistical inference engine to enable investigators in a wide variety of fields to make use of large (terabyte) databases to confront and test physical models. The system targets spatially-mapped heterogeneous data, typical of remote sensing problems. These problems often require simultaneous interpretation of current and archival databases or data collected by multiple sensors with different spatial resolution, type, and sensitivity. For example, researchers in geological,forest and natural resource management, atmospheric, and biological/environmental sciences are using both space- and Earth-based observation networks to accumulate large databases of spatially global information of many types. Similarly, astronomers are mapping large regions of the sky to compile catalogs of sources at different wavebands and differing positional accuracy. The system is designed to free the scientist from the mechanics of large-scale statistical analysis and to make the knowledge obtained at each stage available to direct the ongoing exploration.