论文信息 - Using support vector machines to automatically extract open water signatures from POLDER multi-angle data over boreal regions

Using support vector machines to automatically extract open water signatures from POLDER multi-angle data over boreal regions

This study used support vector machines to classify multiangle POLDER data. Boreal wetland ecosystems cover an estimated 90 x 10/sup 6/ ha, about 36% of global wetlands, and are a major source of trace gases emissions to the atmosphere. Four to 20 percent of the global emission of methane to the atmosphere comes from wetlands north of 40/spl deg/N latitude. Large uncertainties in emissions exist because of large spatial and temporal variation in the production and consumption of methane. Accurate knowledge of the areal extent of open water and inundated vegetation is critical to estimating magnitudes of trace gas emissions. Improvements in land cover mapping have been sought using physical-modeling approaches, neural networks, and active-microwave, examples that demonstrate the difficulties of separating open water, inundated vegetation and dry upland vegetation. Here we examine the feasibility of using a support vector machine to classify POLDER data representing open water, inundated vegetation and dry upland vegetation.

[1] F. Hall,et al. Physically based classification and satellite mapping of biophysical characteristics in the southern boreal forest , 1997 .

[2] Mingkui Cao,et al. Global carbon exchange and methane emissions from natural wetlands : Application of a process-based model , 1996 .

[3] T. Christensen,et al. Vascular plant controls on methane emissions from northern peatforming wetlands. , 1999, Trends in ecology & evolution.

[4] David E. Knapp,et al. Classification of Forest Land Cover in BOREAS. II: Comparison of Results from a Sub-pixel Scale Physical Modeling Approach and a Training Based Method , 1997 .

[5] Claude R. Duguay,et al. Comparison of Evidential Reasoning and Neural Network Approaches in a Multi-source Classification of Alpine Tundra Vegetation , 1996 .

[6] Jon Atli Benediktsson,et al. Neural Network Approaches Versus Statistical Methods in Classification of Multisource Remote Sensing Data , 1989, 12th Canadian Symposium on Remote Sensing Geoscience and Remote Sensing Symposium,.

[7] Guoqing Sun,et al. Mapping of boreal forest biomass from spaceborne synthetic aperture radar , 1997 .

[8] Vladimir Naumovich Vapni. The Nature of Statistical Learning Theory , 1995 .

[9] Charles L. Walthall,et al. Evidence of hot spot directional signature from airborne POLDER measurements , 1997, IEEE Trans. Geosci. Remote. Sens..