标题：Sub-pixel and Maximum Likelihood Classification of Landsat Etm+ Images for Detecting Illegal Logging and Mapping Tropical Rain Forest Cover Types In Berau, East Kalimantan, Indonesia
期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2004
卷号：XXXV Part B7
页码：933-942
出版社：Copernicus Publications
摘要：The tropical forest is depleting at a fast rate due to deforestation and degradation. Illegal logging was reported to be the cause of 50% of the deforestation. Illegal logging is a very pressing issue in Indonesia that is threatening the sustainability of forest management. The detection of the single felling tree which can be characterized as a specific type of illegal logging can provide information for the assessment of related Criteria and Indicator (C&I) of Sustainable Forest Management (SFM) and therefore support the certification of Sustainable Forest Management. This study aims to detect single tree felling in the tropical forest using Landsat-7 ETM+ satellite data and two types of classifiers i.e. maximum likelihood classifier and the sub-pixel classifier. Furthermore, it aims to assess the output of the first objective to support SFM through evaluation of specific C&I. Field data of new logged points representing single tree felling was collected during fieldwork in East Kalimantan, Indonesia in September 2003. The Landsat image was classified using maximum likelihood and sub-pixel classification. The results showed that the accuracy of the sub-pixel classification was higher than the maximum likelihood classification of the 30 m resolution image with an overall accuracy and kappa of 89% and 0.75 versus 79 % and 0.57 respectively. Consequently, more accurate detection of single tree felling can be achieved using the sub-pixel classifier and Landsat-7 ETM+ image. The extracted information can be characterized as planned or illegal with the use of GIS and expert knowledge which helped to identify specific indicators of SFM related with illegal single tree felling. The measurement of these indicators will ultimately support the SFM assessment
关键词：Su-pixel; Maximum Likelihood; Forest; Detection
