Bethel, James
4 CONCLUSIONS
The increased spectral resolution afforded by hyperspectral sensors such as HYDICE, together with appropriately
designed statistical pattern recognition algorithms, can yield excellent and reliable material and land cover region
delineation. Proper modeling of the atmospheric turbulence driven variability in the host aircraft position and angular
trajectory can yield accurate transformations between image and ground coordinates. The combination of these two
approaches to image exploitation, which have traditionally been carried out in isolation, provides unique information
toward the rigorous development of 3D databases in which the features and textures are both correctly labeled and
correctly positioned.
ACKNOWLEDGMENTS
The work described in this paper was sponsored by the U.S. Army Research Office under Grant Number
DAAHO04-96-1-0444. The authors would like to acknowledge the support of Dr. Edward Mikhail, the project director.
They would also like to thank Hank Theiss for his help in data analysis.
REFERENCES
1. Amini, A., Weymouth, T., Jain, R., "Using dynamic programming for solving variational problems in vision",
IEEE Transactions on Pattern Analysis and machine Intelligence, Vol. 12, No. 9, 1990.
2. Landgrebe, D., 1999, "Information Extraction Principles and Methods for Multispectral and Hyperspectral Image
Data”, Chapter 1 of Information Processing for Remote Sensing, edited by C.H. Chen, World Scientific Publishing
Co., River Edge, NJ
3. Lee, C., Theiss, H., Bethel, J., Mikhail, E., 2000, "Rigorous Mathematical Modeling of Airborne Pushbroom
Imaging Systems", Photogrammetric Engineering and Remote Sensing, Vol. 66, No. 4, April 2000, pp. 385-392.
4. Mikhail, E., and McGlone, C., 1980, *Current Status of Metric Reduction of (Passive) Scanner Data", Proceedings
of ISPRS Congress, Commission III, WG III-1, p.504. Hamburg
190 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
p - pd re MN d
— MN PA IAS
