Toward Automatic Cartographic Feature Extraction

Abstract

Over the last twenty years researchers in the field of computer vision have focused on a variety of application areas in order to test their ideas and approaches. The diversity of application areas such as industrial inspection, material handling, autonomous navigation, and physical modeling and recognition, to name a few, illustrates the ubiquitous nature of computer vision. Our research has focused in another application area, that of digital mapping, which is the automation of the cartographic process. Cartography is a human endeavor that attempts to accurately depict our three-dimensional world using two-dimensional media. In modern times the utilization of airborne and space-based remotely sensed imagery has changed the mapping community in a fundamental way. No longer relying entirely on surveyed ground measurements, remotely sensed imagery provides a synoptic view of the environment using a variety of sensors including multi-spectral scanners, black and white photography, color infra-red cameras, and imaging radars.