摘要
洪水是最具破坏性的自然灾害之一。近几十年来,由于全球变暖和大规模的城市化,城市洪水灾害变得更加频繁,且令人们难以预测和防范,危害性日益加剧。遥感技术具有快速、大空间范围、低成本和高效的对地探测能力,在洪水灾害监测中具有巨大的应用潜力。由于城市建筑(如阴影、道路和其他黑暗物体)等低反照率地物的影响及城市景观的复杂性造成难以精准提取城市洪水。因此准确提取城市洪水信息对于监测城市洪水灾害具有重大现实意义。因此,利用Landsat 8 0LI影像,以2016年武汉城市洪水灾害为例,选取城市洪水泛滥区的"湖泊""河流""湖泊河流交汇处"3个区域为典型样区,选用归一化差值水体指数(NDWI)、改进型归一化差值水体指数(MNDWI)、增强型水体指数(EWI)、新型水体指数(NWI)、修订型归一化水体指数(RNDWI)和利用Band1与Band7归一化差分水体指数(NEW)等6种常见水体指数,以大津阈值法估算水与非水地物的分割阈值,对城市洪水范围进行提取,并将洪水提取结果和支持向量机(SVM)监督分类结果对比分析。结果表明,在城市河流洪水泛滥区提取水体效果好的方法依次为EWI> NWI> RNDWI;在城市湖泊洪水泛滥区提取水体效果好的方法依次为EWI> MNDI>NWI;在城市河流湖泊交汇区提取水体效果好的方法依次为EWI> NWI> RNDWI。论文的研究成果为利用遥感影像监测和评估城市洪水灾害提供了重要参考。
Floods are one of the most devastating natural disasters. In recent decades,urban flooding has become more frequent. It is hard to predict and prevent,and the harmful is increasing. Remote sensing technology has great potential for application in flood disaster monitoring because of its rapid,large spatial range,low cost and high-efficiency ground detection capability. Due to the influence of low able do features such as shadows,roads and other dark objects and the complexity of urban landscapes,causing difficulty in accurately extracting urban floods. Therefore,accurate extraction of urban flood information is of great practical significance for monitoring urban flood disasters. Therefore,the study uses the Landsat 8 0 LI image,taking the urban flood disaster in Wuhan in 2016 as an example,and selects three areas of"lake""river"and"intersection of lake and river "as typical sample areas,the flood range was extracted using six common water indices NDWI,MNDWI,EWI,NWI,RNDWI,NEW and estimation of the segmentation threshold of water and non-water features with the Otsu threshold method,andcompare the flood extraction results with the supervised classification results. The results show that the better way to extract water bodies in river areas is EWI > NWI > RNDWI;better way to extract water in the lake research area is EWI > MNDI > NWI;the better way to extract water bodies in the river and lake intersection research area is EWI > NWI >RNDWI. This study provides an important reference for remote sensing image monitoring and assessment of urban flooding.
作者
李昕悦
张锦
贾笑怡
王旭红
LI Xing-yue;ZHANG Jin;JIA Xiao-yi;WANG Xu-hong(College of Urban and Environmental Sciences,Northwest University,Xi'an 710127,China;Shaanxi Key Laboratory of Surface System and Environmental Bearing Capacity,Xi'an 710127,China;Urban Ecosystem Positioning Observation and Research Station,Xi’an 710127,China)
出处
《地下水》
2020年第5期179-183,共5页
Ground water
