James A. Shine

Shine, James A. and Carr, Daniel B.(2003), A New Imagery Classification Method Using Spatial Covariance Information, Computing Science and Statistics, 35, I2003Proceedings/ShineJames/ShineJames.paper.pdf

A New Imagery Classification Method Using Spatial Covariance Information
James A. Shine, (George Mason University), jshine1@gmu.edu, and
Daniel B. Carr, (George Mason University), dcarr@gmu.edu


Classical and modern statistical methods offer a wide variety of approaches to classification of data in general and classification of imagery in particular. None of these approaches explicitly use spatial information. Spatial covariance structures have been used for data prediction, but not directly for classification. This paper describes a classification method using the spatial covariance information in imagery to directly classify images in a supervised approach. A series of thresholds are measured with training data for each class, and a model is then fitted. Each pixel is measured for its fit for each class, and the class with the best fit is chosen. A framework is also described for using multiple bands of information and classifying from the combined bands. Results from multispectral imagery classification will be discussed and analyzed.

Take me back to the main proceedings page.