LANDSAT-BASED LAND COVER CHANGE ANALYSIS IN NORTHEASTERN OREGON'S TIMBER-RESOURCE-DEPENDENT COMMUNITIES

摘要

The necessity for the development of repeatable, efficient and accurate monitoring of land cover change is paramount to successful management of our planet's natural resources. This study provides a generalized framework by which a range of land cover analyses can be accomplished. Through the creation of atmospheric correction and topographic normalization models, and the development of an object-based land cover change rule set, this study explores the accuracy of a range of segmentation parameters and a new principal component analysis (PCA) change detection method. The results suggest that segmentation with a relatively small scale parameter and complete influence of color over shape produced the most accurate classification. Additionally, the PCA change detection method appears to be an effective predictor of change, although further comparative analyses are needed.