Wall-to-wall Landsat TM classifications for Georgia in support of SAFIS using FIA plots for training and verification

Abstract

Wall-to-wall Landsat TM classification efforts in Georgia require field validation. Validation uslng FIA data was testing by developing a new crown modeling procedure. A methodology is under development at the Southern Research Station to model crown diameter using Forest Health monitoring data. These models are used to simulate the proportion of tree crowns that reflect light or on a FIA subplot bass. The subplot crown proportions are averaged and compared to Landsat TM classifications for verification purposes. Resolution differences between field data and Landsat TM data make comparisons challenging. Positive correlations between the two types of data were recorded for 4 of the 5 FIA plots tested. Differences on the 5^th plot may be contributed to mis-registration of the two data sources or mis-classfication of the TM imagery.

Citation

Cooke, William H., III; Hartsell, Andrew J. 2000. Wall-to-wall Landsat TM classifications for Georgia in support of SAFIS using FIA plots for training and verification. Proceedings of the Eighth Forest Service Remote Sensing Applications Conference