Computer Science Department
School of Computer Science, Carnegie Mellon University


Automated Cartographic Feature Attribution Using
Panchromatic and Hyperspectral Imagery

David M. McKeown, Jr., Chris McGlone, Steven Douglas Cochran,
Wilson A. Harvey, Jefferey A. Shufelt, Daniel Yocum

September 1998

This paper was presented at the
1998 DARPA Image Understanding Workshop
20-23 November 1998, Monterey, California.

Available by request only at Digital Mapping Laboratory Publications Archive

Keywords: Automated cartographic feature extraction, automated road detection, building detection, data fusion, feature aggregation, HYDICE sensor system, knowledge-based systems, multispectral analysis, performance evaluation, scene registration, spatial databases, stereo, surface material classification, synthetic environments

This technical report describes research in the automated analysis of hyperspectral imagery (HYDICE). The goal of our research is to investigate issues in automatic surface material mapping using 210 channel image data from an airborne scanner sensitive from the visible through shortwave infrared wavelengths. Research issues addressed include sensor modeling, geometric correction and positioning, material classification experiments using two different interpretation models, and the fusion of geometric information from high resolution panchromatic imagery. Evaluation results are presented for building and road detection and attribution. The use of the resulting material classification maps for visual simulation is also presented.

22 pages

Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by