Object-Based Segmentation and Classification of One Meter Imagery for Use in Forest Management Plans

Publication Year:
2010
Usage 6046
Downloads 5914
Abstract Views 132
Repository URL:
https://digitalcommons.usu.edu/etd/653
Author(s):
Wells, W. Kevin
Tags:
Forest Management Plan; GIS; GIS Model; Python Script; Vegetation Types; Remote Sensing; Forestry and Wildlife; Agricultural and Resource Economics; Forest Management; Geographic Information Sciences
thesis / dissertation description
This research developed an ArcGIS Python model that extracts polygons from aerial imagery and assigns each polygon a vegetation type based on a modified set of landcover classes from the Southwest Regional Gap Analysis Project. The model showed an ability to generate polygons that accurately represent vegetation community boundaries across a large landscape. The model is for use by the Utah Division of Forestry, Fire, and State Lands to assist in the preparation of forest management plans. The model was judged useful because it was easy to use, it met a designated 50% threshold of useable polygons, and it met a designated 50% threshold of vegetation class assignment accuracy.