《Wetlands》杂志刊登“利用多时相反射率方法进行退化森林湿地分类”
作者:Richard F. Keim, Jason A. Zoller, DeWitt H. Braud, Brandon L. Edwards
刊物:《Wetlands》,2013年10月
关键词:沼泽;可变性;多变量统计;植物
摘要:湿地生态系统的遥感分类可能会因为植被和水文条件的时空变异而变得困难,我们开发了一种多时相分类方案,通过在多时相反射率的主要成分进行像素分类,采用landsat-7三个波段的数据,定义分类的反射率的主要成分和路易斯安那野外场地的植被状况呈现统计相关性。因此分类结果表示所需的生态梯度比单图像分类更有效。
Classification of Forested Wetland Degradation Using Ordination of Multitemporal Reflectance
Authors: Richard F. Keim, Jason A. Zoller, DeWitt H. Braud, Brandon L. Edwards
Journal: Wetland, December 2013
Key words:Swamp;Variability;Multivariatestatistics;Baldcypress
Abstract: Remote sensing classifications of wetland ecosystems can be difficult because of temporal variability of plant cover and hydrological conditions. We developed a multitemporal classification scheme to control for these factors in forested wetlands by classifying pixels within a principal-components ordination of multitemporal reflectance, using data from three Landsat bands from each of seven scenes across a 3-year period in each of two phenological conditions. The principal components of reflectance that defined the classification were statistically related to vegetative conditions of the Louisiana field sites in interpretable ways, so the resulting classification represents the desired ecological gradients more robustly than single-image classifications.
原文链接:http://link.springer.com/article/10.1007/s13157-013-0466-1
翻译:刘金轲;审核:翟正丽