摘要
子像元分解分类是遥感领域比较棘手的方法,尽管目前有不少软件和算法可以进行这方面的处理,但对其结果的应用还处在初级探索阶段。通过回顾目前用于解决像元分解问题的基本理论和方法,介绍了当今用于子像元分解分类的3种技术方法(即ILWIS3.0子像元线性分解分类、Erdas8.5子像元分解分类、eCognition3.0最近邻分解分类)的基本原理、算法、参数设定和计算过程,以NOAA影像的第一、第二波段数据(卫星扫描时间是1998年8月11)对江西鄱阳湖的洪水监测为例,用上述3种软件和方法进行了实际分解分类计算,并在文中展示了3种方法的最后分类结果。其次对3种分类方法的基本原理、设定参数、适合应用和各自存在缺点进行了理论、应用上的分析对比,还就子像元分解分类结果(每个像元含有子像元的百分比)的应用(即子像元的空间分配问题)提出了个人意见,以期能为今后遥感同仁在以后的类似具体问题上提供一点参考价值。
Sub-pixel classification is a tough issue in remote sensing field, the application of its calculation result is at primary, probation stage although many kinds of software and algorithm can be used to process this problems. After retrospecting the basic theory and methods of dealing with unmixing pixels, this paper made an introduction to the principles, algorithms, parameter setting and computing process of three Sub-pixel calculation methods (those are ILWIS3.0's Linear Unmixing, Erdas8.5's Sub-pixel Classifier, eCognition3.0's Nearest Neighbor); conducted a case study of flood monitoring in Poyang Lake region with image data (scan time is August 11, 1998) of channels 1, 2 of NOAA AVHRR using the aforementioned three kinds of software and methods, the classified results are displayed in this paper. Second, a theoretic, practical and analytic comparison study was made of these three software and Sub-pixel calculation methods in the aspects of Basic principles, Parameters to be set, Application fields and their respective drawbacks. Finally recommended was the author's opinion on the use (spatial assignment of sub-pixels of each pixel) of the calculation results (percentage of sub-pixels in each pixel) with these three Sub-pixel calculation methods in a hope to provide some reference to future similar application study to the RS scientists.
出处
《遥感技术与应用》
CSCD
2005年第2期266-271,共6页
Remote Sensing Technology and Application
基金
荷兰"地球监测和地球信息科学国际研究院(ITC)"研修"地球观测和地球信息科学在自然灾害研究中的应用"方向期间的部分研究成果
人事部"留学人员科技活动项目择优资助经费"(国人部发〔2003〕50号)的支持。
关键词
子像元方法
案例研究
应用对比
Sub-pixel Methods, Case study, Comparison study
