Forest/Nonforest Classification of Landsat TM Data For Annual Inventory Phase One Stratification

Abstract

Launch of Landsat 7 creates the opportunity to use relatively inexpensive and regularly acquired land cover data as an alternative to high altitude aerial photography. Creating a forest/nonforest mask from satellite imagery may offer a cost-effective alternative to interpretation of aerial photography for Phase One stratification of annual inventory plots. This paper describes the procedures: they include image rectification, registration, and spatial filtering to allow accurate co-location with field plots and attempt to compensate for minor plot location errors. Identification of clouds and their removal from further analysis is outlined. Image alarms are described as a coarse filter for arriving at a forest/nonforest mask, with unsupervised classification as the fine filter. Accuracy assessment results for single-date, dual-date, filtered and unfiltered combinations are reported, as well as cost estimates.

Citation

Rack, Jim. 2001. Forest/Nonforest Classification of Landsat TM Data For Annual Inventory Phase One Stratification. In: Reams, Gregory A.; McRoberts, Ronald E.; Van Deusen, Paul C., eds. 2001. Proceedings of the second annual Forest Inventory and Analysis symposium; 2000 October 17-18; Salt Lake City, UT. Gen. Tech. Rep. SRS-47. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. pp. 8-10