Estimation of timber volume by LANDSAT/TM imagery

摘要

The objective of this study was to evaluate the accuracy of timber volume estimates made in three forest types (plantations dominated by Japanese larch, natural broadleaf forest dominated by Quercus crispula, and natural broadleaf forest dominated by other species) at the Kyushu University forest in Hokkaido. Estimates were made using the weighted k–nearest neighbor (k–NN) method, combining inventory data with satellite image data and environmental factors. The estimates were then compared to true timber volumes obtained from CFI data. In the plantation areas, the degree of accuracy was improved as environmental factors were added. In the natural forest areas, environmental factors had no effect on estimate accuracy.