The Austin History Center, the City of Austin and UT Libraries maintain substantial collections of scanned maps, which are needed for historical reference when conducting research or planning new projects. Unfortunately, these scanned maps can be difficult to use due to challenges involved with discovering, processing, and analyzing map images in their current digitized form. Our team is acutely aware of how collections of map images could be further developed to make them easier to work with, but face significant challenges in improving these resources due to the large number of maps and the significant amount of time it takes to process them. UT Libraries has explored a number of methods for increasing the speed of map processing including hiring student assistants to focus on this work, pursuing funding to support a commercial crowd-sourced georeferencing platform, and developing scripted processes to analyze map images and enhance their research potential. This project will build on this initial exploration with expert technical guidance and input from researchers and the City of Austin. The team will investigate image processing, computer vision, machine learning, optical character recognition, and GIS approaches to determine which have the highest likelihood of success in making map data easier to search and more accessible to general users. The project strives to bring research from lab to application, making those who rely on this mapping data more effective.
Aaron Choate (University of Texas Libraries), Michael Shensky (UT Libraries and Department of Geography and the Environment), Danna Gurari (School of Information), William Doolittle (Department of Geography and the Environment), Ross Clark (City of Austin Water), Jennifer Hecker (City of Austin History Center) and Mike Miller (Cityof Austin History Center)