FAST FOREST MONITORING ALGORITHM FOR LANDSAT TM/ETM+ IMAGE DATA

摘要

Forests are vital earth ecosystems that provide a variety of ecological, social and economic values. Deforestation and forest degradation, however, have made forests vulnerable, causing significant impacts on global climate. Forest resources in Vietnam are decreasing dramatically, especially in Kon Turn province which is located in Central Highlands of Vietnam. Quantifying spatial forest change information is important for assessing influences of human activities and environmental changes on ecosystem sustainability. This is a report on development of new forest monitoring algorithm using Landsat TM and ETM+ dataset. The algorithm allows fast forest classification and change detection. The algorithm was developed based on usage shape of spectral reflectance curve. Spectral reflectance curve of similar ground objects is encoded using modulation of the curve and using the modulation of spectral reflectance curve forest can be fast and almost automatically extracted from the Landsat TM and ETM+ image. An area of Kon Turn province of Vietnam has been used as test site. Validation was made by field work conducted in 2012 and comparison with latest land use map.