Title :
Online, map-based estimation of forest attributes
Author :
McRoberts, Ronald E. ; Miles, Patrick D.
Author_Institution :
USDA Forest Service, North Central Res. Station, St. Paul, MN, USA
Abstract :
Areal estimates of forest attributes are obtained from maps constructed using forest inventory plot data, satellite imagery, and the k-nearest neighbors technique. The map-based areal estimates are calculated by aggregating attribute predictions for individual image pixels and are statistically comparable to estimates obtained for the same areas using inventory plot data only. Online, map-based estimation provides general access to information derived from inventory data without revealing exact plot locations or proprietary ownership information, two crucial issues for forest inventory in the USA. Techniques are discussed for increasing the speed of online estimation without compromising accuracy.
Keywords :
Internet; cartography; forestry; geographic information systems; image resolution; forest attributes; forest inventory plot data; image pixels; information access; k-nearest neighbors technique; map-based areal estimation; proprietary ownership information; satellite imagery; Conferences; Global Positioning System; Image databases; Large-scale systems; Pixel; Prototypes; Sampling methods; Satellites; State estimation; US Department of Agriculture;
Conference_Titel :
Database and Expert Systems Applications, 2005. Proceedings. Sixteenth International Workshop on
Print_ISBN :
0-7695-2424-9
DOI :
10.1109/DEXA.2005.149
