DIGITIZING THE COLLECTIONS OF THE PERKINS GEOLOGY MUSEUM TO SUPPORT ON-LINE LEARNING APPLICATIONS ABOUT VERMONT GEOLOGY

With funding from the Institute of Museum and Library Services (IMLS), we are digitally archiving the holdings of the Perkins Geology Museum at the University of Vermont. The new digital archive houses images of fossils, rocks, minerals, maps, thin-sections, 35mm slides, photographs, and field outcrops. In addition, the archive incorporates several "digital-ready" sets of images from the museum's Landscape Change Project and Holocene Sediment Cores Project. The final archive will house an estimated 40,000 images from 20,000 objects. Access to the archive is twofold, via a Voyager library catalog (http://perkinscatalog.uvm.edu) and a web site (http://www.uvm.edu/perkins/). Educators, researchers, and students will work with pedagogical "hints for archive use," on-line curricula, and on-line "collection" and "viewing" tools. Images of Vermont specimens are linked hierarchically to thin-sections, chemical composition files, outcrop location images, and/or map images where appropriate, creating a deeper context for select images.

Our digitization sequence includes the following tasks: Confirm existing museum catalog data, capture digital image, verify image and add metadata, send to master MySQL database, create Voyager record, create XML file for the National Science Digital Library (NSDL), and obtain external expert feedback about unusual specimens. The main MySQL database stores collection data, image metadata, annotations and labels for images, and links to associated images. The Voyager catalog provides the metadata in USMARC format where records are Z39.50 compatible. Both the MySQL database and the Voyager catalog use standard GeoRef Thesaurus terminology to describe and organize images. Additional paleontological descriptions reference standard texts such as the Treatise on Invertebrate Paleontology and Index Fossils of North America. Web pages are authored in Hypertext Markup Language (HTML) and PHP: Hypertext Preprocessor (PHP).

Currently, we store uncompressed TIFF files (1600x1200 pixels, 5-6MB), compressed PNG files (1600x1200 pixels, 2-3MB) and compressed JPEG files (1024x768 pixels, 100K). Our goal is to revert to a single, resizable PNG file (3000x2000 pixels, 7MB) and forgo the necessity of storing JPEG images for on-line use.