An empirical study of image processing methods for land cover classification and forest cover change detection in Northeastern Oregon's timber resource-dependent communities (1986-2011)

A study was performed to evaluate remote sensing methods for classifying land cover and land cover change throughout a two-county area in Northeastern Oregon (1986-2011). In the past three decades, this region has seen significant changes in forest management -- changes that can be readily identified from the synoptic perspective. This study employs an accuracy assessment-based empirical approach to test a number of advanced digital image processing techniques that have recently emerged in the field of remote sensing. The accuracies are assessed using traditional and area-based error matrices. It was determined that, for single-time land cover classification, Bayes pixel-based classification using samples created with segmentation parameters of scale 8 and shape 0.3 resulted in the highest overall accuracy. For land cover change detection, it was determined that Landsat 5 TM band 7 with a change threshold of 1.75 SD resulted in the highest accuracy for forest harvesting detection.