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：2000
卷号：XXXIII Part B7(/1-4)
页码：876-885
出版社：Copernicus Publications
摘要：Global carbon emissio ns leading to change in climate and ecology systems have pro mpted the international co mmunity to establish a series of agreements aimed at reducing such emission. Australia as a signatory country to the 1997 Kyoto Protocol, is currently establishing a National Carbon Accounting System, to estimate and monitor carbon emissions from all major sectors. Forest and woodland biomass mapping, using remotely sensed data, can provide a unique, efficient and reliable assessment for biomass stocks and changes. Studies internationally have recognised that, amo ng all the remote sensing technologies, SAR has the greatest potential to quantify biomass and structural diversity because of its ability of penetration. While SAR data are promising for biomass mapping, SAR image classification based on pixel-analysis suffers from the existing speckle problem. It has been suggested that the SAR data analysis be best based on area-analysis. Segmentation is the process of partitioning an image into uniform areas and providing a base for area-analysis. In this paper, a segmentation method using the Gaussian Markov rando m field model, developed by authors in recent years is used to segment SAR images
