Exploiting Spatial Information for Improving Statistical Classification Methods and Error Analysis: Applications to Satellite Remote Sensing and Vegetation Mapping

Authors

Dr. Brian Steele, Department of Mathematical Sciences, University of Montana, Missoula and candidate for the Statistics position

Document Type

Presentation Abstract

Presentation Date

3-1-1999

Abstract

Satellite remote sensing has become an important tool in natural resource assessment and management during the past 10 years. However, the sheer quantity of data contained satellite imagery poses substantial difficulties for extracting usable and accurate information. Mathematics, and statistics in particular, have not been utilized to their full extent for analyzing satellite imagery, and so the area is rich in opportunities for statistical research. In this talk, new approaches to statistical classification and analysis of classification errors based on exploiting spatial information are discussed. Ongoing efforts aimed at assessing habitat suitability of the Salmon-Selway Wilderness for grizzly bears will be mentioned. The talk is intended to be accessible to a broad audience.

Additional Details

Monday, 1 March 1999
4:10 p.m. in MA 109
Coffee/Tea/Treats 3:30 p.m. in MA 104 (Lounge)

Recommended Citation

Steele, Dr. Brian, "Exploiting Spatial Information for Improving Statistical Classification Methods and Error Analysis: Applications to Satellite Remote Sensing and Vegetation Mapping" (1999). Colloquia of the Department of Mathematical Sciences. 33.
https://scholarworks.umt.edu/mathcolloquia/33