|
Track: Database Design and Automation
Brian Graff
U.S. Army
CETEC-TD-TD
7701 Telegraph Road
Alexandria, VA 22310
Telephone: 703-428-6071
Fax: 703-428-6425
E-mail: bgraff@tec.armay.mil
A Semiautomated Approach Extracting Roads from Scanned Color Maps
Data used in a GIS commonly consists of various attributed vector data layers. Unfortunately, this type of data is not always available for many areas of the world or at all scales required for specific types of analyses. One method of generating digital vector data is to extract the information from color topographic maps. These maps are rich in a variety of features that are desirable for use in a GIS. The traditional approach to this type of extraction is to digitize the required features manually from either hardcopy or scanned color maps, which is labor intensive and time-consuming. Thus, the U.S. Army Topographic Engineering Center (TEC) is investigating semiautomated methods to extract and attribute roads from scanned color topographic maps. The first step in this process is to extract a preliminary classification of a road network using a commercial off-the-shelf software package that applies a neural network approach. This results in a raster file containing a rough road network and numerous
nonroad artifacts such as speckles, sporadic text, and so forth. TEC has written ArcInfo ARC Macro Language (AML) scripts to eliminate the artifacts and yet retain the roads by using knowledge-based geometric descriptors. Roads are then attributed by applying a set of rules about known road characteristics. The goal is to use a semiautomated approach to extract and attribute a vector transportation layer suitable for use in a GIS.
|
|
