多时相Landsat影像结合地形和水文信息进行洪泛区分析  

Floodplain Analysis by Using Multi-temporal Landsat Imagery Combined

作　　者：陈隽敏[1] 

机构地区：[1]中国能源建设集团广东省电力设计研究院有限公司,广州510663

出　　处：《遥感信息》2015年第3期74-79,共6页Remote Sensing Information

摘　　要：洪泛区是地球上最丰富生物栖息地之一,具有很高保育价值。由于其分布广、具有季节性河水泛滥特点,难以通过短期监测方法进行获取,可通过系统地分析流域内水文观测站的日流量信息,选取洪水期和枯水期Landsat卫星影像,作为洪泛区分析的影像分析源。该文依据洪泛区域在近红外和中红外的光谱特征和地形特点,建立多时相洪泛区分析模型,通过非监督分类方法及子类分类进行洪泛区信息提取。实践证明,该模型是一种比较经济、可靠和高效的洪泛区信息提取方法。Floodplain is one of the richest habitats on earth with a high conservation value.Because of its wide distribution and seasonal flooding characteristics,it is difficult to identify floodplain through short-term monitoring methods.By systematic analysis on daily flow gages data in the watershed,this paper selected a pair of representative high flow and low flow Landsat satellite images as data source.Based on floodplain spectral characteristics in the near-infrared and mid-infrared band,combined with its topographic features,it used unsupervised classification method and sub-unsupervised classification to extract floodplain information.The practice has proved that the model is more economical,reliable and efficient for floodplain identification.

关 键 词：洪泛区 水文观测站 LANDSAT 山体阴影 变化检测 分层样例抽取 

分 类 号：P332[天文地球—水文科学]