| Título: | Reflections on the estimation of stand-level forest characteristics using Landsat satellite imagery |
| Autores: |
Hojung Kim; University of Georgia Pete Bettinger; University of Georgia Chris Cieszewski; University of Georgia |
| Fecha: | 2012-08-17 |
| Publicador: | Applied Remote Sensing |
| Fuente: |
 |
| Tipo: |
Literature Review |
| Tema: |
forestry; remote sensing Remote sensing, forest inventory, regression, nearest neighbor, image classification land classification |
| Descripción: | Considerable attention has been applied to the use of Landsat satellite imagery in estimating forest conditions. Here, we focus on the stand-level forest characteristics that have been estimated with this imagery, the classification techniques that have been employed, and the ancillary data that have been used to assist in the process. Based on the peer-reviewed research we located, some gaps in the literature concerning image classification techniques remain. With regard to the algorithms employed in the image classification processes, various forms of regression analysis seem to be the most often used techniques, while the k-Nearest Neighbor (kNN) technique has been increasing in value, yet other classification techniques (e.g., kriging, neural networks) have only begun to be explored and may have value in some situations. In terms of specific forest conditions, the use of Landsat satellite imagery for estimating above ground biomass has been heavily investigated, and opportunities continue to exist for refining classification techniques aimed at the classification of forests by discrete age classes, by contiguous species groups, and by height classes. |
| Idioma: | Inglés |