At the Maya archæological site of Sayil, Berkeley archæologists are using multispectral image analysis tools to discover the complex spatial patterns of prehistoric community organization.

   

In this example, nine layers of ceramic form data (e.g. bowls, jars, plates, etc.) have been subjected to unsupervised classification using a neural network algorithm.  The resulting ten classes indicate areas having the same ceramic assemblage--areas likely to have been the locations of similar activities.

Webpage prepared May 1999 for the UCGIS Congressional Reception.