Computer Science Department
School of Computer Science, Carnegie Mellon University


Preliminary Results on the Analysis of HYDICE Data for Information Fusion in Cartographic Feature Extraction

Stephen J. Ford, Dirk Kalp, J. Chris McGlone, David M. McKeown, Jr.

June 1997

This paper was presented at the SPIE Conference on "Integrating Photogrammetric Techniques with Scene Analysis and Machine Vision III", April 21-23, 1997, held in Orlando, Florida. An identical text version, without color figures, appears in the Proceedings of the Conference, SPIE Volume 3072.

Unavailable Electronically

Keywords: HYDICE, airborne hyperspectral, linear pushbroom sensor, photogrammetry, cartographic feature extraction, surface material classification

During October 1995, high spatial resolution hyperspectral imagery was acquired with the Naval Research Laboratory's (NRL) HYDICE sensor system over Fort Hood, Texas. Fort Hood, Texas has been a focal point for our research and experimentations in automated cartographic feature extraction, stereo analysis and spatial database construction using panchromatic mapping photography and digital cartographic datasets. With the addition of a high resolution hyperspectral dataset, comparable in spatial resolution to panchromatic mapping imagery, opportunities exist to exploit the inherent spectral information of the hyperspectral imagery to aid urban scene analysis for cartographic feature extraction and spatial database population.

This paper discusses our efforts in registering HYDICE imagery, utilizing HYDICE hyperspectral imagery for surface material/land cover classification and merging this information in a common geographic framework with individual datasets or outputs of our cartographic feature extraction systems. Test areas selected from the Fort Hood image and cartographic dataset will illustrate this process flow.

21 pages

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

This page maintained by