Date of Award

Fall 2012

Project Type


Program or Major

Natural Resources

Degree Name

Master of Science

First Advisor

Russell G Congalton


The use of satellite imagery to classify New England forests is inherently complicated due to high species diversity and complex spatial distributions across a landscape. The use of imagery with high spatial resolutions to classify forests has become more commonplace as new satellite technology become available. Pixel-based methods of classification have been traditionally used to identify forest cover types. However, object-based image analysis (OBIA) has been shown to provide more accurate results. This study explored the ability of OBIA to classify forest stands in New Hampshire using two methods: by identifying stands within an IKONOS satellite image, and by identifying individual trees and building them into forest stands.

Forest stands were classified in the IKONOS image using OBIA. However, the spatial resolution was not high enough to distinguish individual tree crowns and therefore, individual trees could not be accurately identified to create forest stands. In addition, the accuracy of labeling forest stands using the OBIA approach was low. In the future, these results could be improved by using a modified classification approach and appropriate sampling scheme more reflective of object-based analysis.