We would like to announce the beta release of "starspan". Starspan is designed to bridge the raster and vector worlds of spatial analysis using fast algorithms for pixel level extraction from point and polygon data.

The main purpose of starspan is to generate databases of pixel values (from one or a set of raster images) falling within points or polygons, and fusing them with the database attributes from the vector files. This allows a user to do statistical analysis of the pixel vs attribute data in any existing statistical package (the output of starspan is a CSV file) and can greatly speed up classification training and testing using many image analysis packages (e.g. ENVI, Imagine, PCI) by focusing the classification on only pixels with known attributes, not the entire image.

STARSPAN runs on any Unix/Linux flavor (including MacOs X and Windows via Cygwin) although it has been only fully tested on Linux and MacOS X. The software is completely free. If you are interested in testing/using the software, the main website is located at:

http://starspan.casil.ucdavis.edu/

We encourage feature requests and bug reports, and these can be posted on our gforge website at:

http://casil.ucdavis.edu/projects/starspan/

Sincerely,

Jonathan Greenberg and Carlos Rueda