EN中文

地球信息科学学报 >

2017 , Vol. 19 >Issue 5: 702 - 712

DOI: https://doi.org/10.3724/SP.J.1047.2017.00702

遥感科学与应用技术

最优分区与最优指数联合的水体信息提取

  • 饶萍 , 1, 2 ,
  • 王建力 , 1, *
展开
  • 1. 西南大学地理科学学院,重庆 400715
  • 2. 贵州工程应用技术学院,毕节 5517001
*通讯作者：王建力（1969-）,男,山东威海人,博士,教授,博士生导师,研究方向为自然地理学。E-mail:

作者简介：饶 萍（1979-）,女,贵州毕节人,博士生,研究方向为遥感图像处理及应用。E-mail:

收稿日期: 2016-11-23

  要求修回日期: 2017-01-23

  网络出版日期: 2017-05-20

基金资助

中央高校基本科研业务费专项资金项目（XDJK2011C089）

国家自然科学基金项目（40971122）

收起

Water Extraction Based on the Optimal Subregion and the Optimal Indexes Combined

  • RAO Ping , 1, 2 ,
  • WANG Jianli , 1, *
Expand
  • 1. School of Geography Sciences, Southwest University, Chongqing 400715, China
  • 2. Guizhou University of Engineering Science, Bijie University, Bijie 551700, China
*Corresponding author: WANG Jianli, Email:

Received date: 2016-11-23

  Request revised date: 2017-01-23

  Online published: 2017-05-20

Copyright

《地球信息科学学报》编辑部 所有
Fold

摘要

单波段阈值法和水体指数法具有方法简单、耗时少的优点,成为水体信息提取常用的方法,但对于面积较大、类型多样、影响因素复杂的区域,在全局使用任何一种方法均无法满足精度要求。为精确提取地表覆盖类型复杂、水体类型多样的山区水体信息,本文分别对归一化水体指数（MNDWI）、自动水体提取指数（AWEI）和归一化三波段指数（NDTBI）3种指数的阈值进行分析,以最优阈值分区方案进行分区并构建3个单指数决策树,寻找不同水体类型的最优指数,按照最优原则重构联合指数决策树来提取水体信息,其中,3种单指数提取法的Kappa系数分别为：0.863, 0.854, 0.862,最优指数联合决策树法的Kappa系数为 0.881。结果表明,基于3个指数构成的最优指数联合决策树的提取方法可以达到最高精度,由此说明,采用该法用于水体信息提取可得到最佳效果。

关键词： 水体; 决策树; 水体指数; 信息提取; 遥感

本文引用格式

饶萍 , 王建力 . 最优分区与最优指数联合的水体信息提取[J]. 地球信息科学学报, 2017 , 19(5) : 702 -712 . DOI: 10.3724/SP.J.1047.2017.00702

Abstract

It is very important to search a methodology to extract surface water quickly, accurately and efficiently. Single band thresholding and water indices are commonly used water extraction methods because of ease of use and the fact that these methods are computationally less time-consuming than alternative approaches. However, in environments where the area is larger, the type of water is various and the influencing factors are complicated, simple classification methods such as two-band water indices and single-band thresholding may not sufficiently and accurately distinguish between water and non-water pixels. In this paper, in order to extract water accurately in the complex mountain area where all the types of water and interference factor exist, the threshold value of three indices is analyzed respectively, that are Moderate Normal Different Water Index (MNDWI), Automate Water Extraction Index (AWEI) and Normal Difference Three Band Index (NDTBI). Then, the zoning scheme based on the optimal threshold value is used to build three decision classifier based on MNDWI, AWEI and NDTBI respectively. After that, comparative analysis is executed among the effect of the three classifiers for different types of surface water. Lastly, according to the optimal principal, a decision tree classifier of three indices (MNDWI, AWEI and NDTBI) united is reconstructed to extracted water. The results show: the Kappa coefficients means of three methods on single index MNDWI, AWEI and NDTBI in three test sites are 0.863, 0.854 and 0.862 respectively. The Kappa coefficients means of the combined indices method is 0.881. Thus, the water extraction method based on the optimal partition and three indices combined possess have the highest accuracy and the best effects. According to the research, we find that the key factor of the precision improved is the optimal zoning method based on the index to achieve hierarchical extraction. No matter which method among the single index or the united indices, the accuracy is all within the permitted range. In comparison of three single indices methods, extraction of region water such as lake and reservoir is suited for the MNDWI, the line water such as river and stream works best with AWEI, the small, narrow river affected by the channel sandstone has good success with NDTBI. However, the combination decision tree classification of the optimal zoning and the optimal method of indices united will be satisfied for water extraction.

Key words： water; decision tree; water index; information extract; remote sensing

1 引言

地表水是至关重要的地球资源之一,随着土地利用/土地覆盖变化、气候变化以及其它环境因素的变化,地表水也随时间和空间发生变化。多年来,地表水的变化对于生态、社会、健康以及经济效应的影响一直是学术界关注的问题[1-7]。快速、准确和高效地提取地表水体具有重要意义。
近年来,各种不同分辨率遥感影像被广泛应用于水体信息提取,许多学者对水体信息遥感处理模型、定量提取方法等方面开展了广泛的研究,取得了影像分类法[8]、线性分解法[9]、单波段阈值法[10]、水体指数法[11-16]、谱间关系法[17]及不同方法的综合集 成[18-21]等众多的研究成果。Billa等[22]利用 AVHRR数据来反演云层表面的亮度温度,同时对云进行分类和等级划分,通过设定经验阈值来确定雨量强度较大的区域,从而对以往的数值雨量预报进行补充,对洪水灾害进行提前预报。闫霈等[23]在分析半干旱地区水系与背景噪音反射特点的基础上,提出了增强型水体指数EWI（Enhanced Water Index）,有效区分了半干涸河道与背景噪音。杨莹等[24]以洪泽湖 Landsat TM 影像为研究对象,综合利用多波段谱间关系TM2+TM3>TM4 +TM5和单波段TM5建立了适合于平原湖泊水域的水体提取方法。汪金花等[25]利用多波段谱间关系法对水体及其它几种物体光谱特性进行了实验分析。由于底面物质成分、深浅、水质成分及其周围环境的影响等因素,即便是同一地域内地表水的波谱特征也是有差异的,在遥感图像上呈现出的影像特征也并不能保持均衡,因此,在全局采用统一模型来提取地表水体信息,难以达到精确提取的要求。鉴于此,本文根据区域地表水体特征及其影响因素,在分类分析的基础上对研究区进行最优分区,并通过不同方法的比较,研究得出各类地表水体信息的最佳指数提取方案,并在全局范围内对不同水体类型区应用最佳指数联合进行提取,从而达到快速、精确提取水体信息的目的。

2 研究区及数据源

本文选取毕节市七星关区（3421 km2）作为研究区（图1）,该区境内山峦重叠,平均海拔1511 m,地貌破碎,土壤类型复杂,河流、湖泊、水库、坑塘等大小水体纵横分布,高原、山地、盆地、谷地、峰丛、槽谷、洼地等交错其间,地表特征复杂,该研究的区域选取具有贵州高原典型性。
Fig. 1 Location of the study area

图1 研究区位置图

注：红色和蓝色标示的矩形区域分别对应图2、图4、图5和表2中的3个测试区域

Fig. 2 The water and non-water noisy in the three test sites

图2 3个测试点大气校正前后的彩色影像及水体解译结果

Fig. 3 Threshold images of AWEInsh and AWEIsh

图3 AWEI指数阈值图像

Fig. 4 Threshold images of MNDWI

图4 MNDWI指数阈值图像

Fig. 5 Threshold images of NDTBI

图5 NDTBI指数阈值图像

针对研究区的实际情况,选取3种不同环境下的区域作为精度评价的测试区域（图1中的3个测试点）,在这3种区域图像上,地表水体受低反射地表、山体阴影及水体底面的沙石等干扰因素的影响。
本文使用的数据为美国Landsat 8 OLI影像数据（数据来源于“中国科学院计算机网络信息中心地理空间数据云”）,景的行列号为128/41,产品类型为L1T,获取日期为2013年7月9日,云的覆盖率为4.76,但多集中分布于影像的左上角区域,研究区内仅有极少量的云朵及其阴影,能够满足水体信息提取研究的要求。
Lansat 8卫星所携带的陆地成像仪（Operational Land Imager,OLI） 具有较高的影像质量,包含8个多光谱波段（30 m）和1个全色波段（15 m）,本文主要使用波段其中的6个波段数据,分别为蓝光波段（0.450~0.515 μm）,绿光波段（0.525~0.600 μm）,红光波段（0.630~0.680 μm）,近红外波段（0.845~0.885 μm）和2个短波红外波段（1.560~1.660 μm,2.100~2.300 μm）。
选用2010年的日本ALOS高分辨率全色影像（2.5 m全色）作为参考数据。

3 研究方法

3.1 影像预处理

本文应用2种方式对以数字形式存储的Landsat 8 OIL影像进行辐射校正。①将原图像的DN值转换到大气顶层反射值,用以创建水体指数图像;②在辐射定标的基础上,应用ENVI 5.0中的FLAASH（Fast Line-of-Sight Atmospheric Analysis of Spectral Hypercubes）大气校正模型进行大气校正,获取地表反射值,用做土地覆盖类型的精确解译。
将参考数据ALOS全色影像与OLI 影像按图像对图像的方式进行配准,对于每一个测试点,手动选取25个控制点进行校正,均方根误差控制在0.5个像元之内。

3.2 影像解译

遥感影像信息提取中,所有模型的设计都是基于对遥感影像的正确解译基础之上,因此,影像的解译至关重要。在原始影像图及大气校正后的影像图上（图2）,面状水体（如水库、湖泊和坑塘等）呈现蓝色、深蓝色,线状水体（河流）主要呈现黑色、蓝色、蓝紫色或紫色。容易与水体造成混淆的因素主要有2种：山体阴影及暗色建筑地表。山体阴影在原始彩色影像上常表现为黑色,容易与水体相混淆,而在大气校正后的影像上山体阴影的影响很小,少部分比较深的阴影呈现墨绿色或者深绿色,通过比较也很容易分辨;而研究区的暗色建筑地表,无论是在大气校正前还是校正后的影像上,都易与水体相混淆,因此可借助高分辨率影像来进行判别。
基于上述分析和准确解译,将大气校正影像作为方法设计的初始数据,将ALOS高分辨率影像解译结果（图2）作为精度评价的验证数据。

3.3 水体指数

当前,在各种水体信息提取方法中,水体指数法因其具有使用简单和计算耗时少的特点,已成为常用的水体提取方法。
目前常用的水体指数主要有以下几种：McFeeters[12]提出基于Lansat TM绿光波段（波段2）和近红外波段（波段4）的归一化水体指数（Normalized Difference Water Index, NDWI）;Rogers 和Kearn[13]提出的基于TM红光波段（波段3）和短波红外波段（波段5）的NDWI;Xu[14]于提出的基于TM波段2和波段5的改进的归一化水体指数（Modified Normalized Difference Water Index, MNDWI）;Feyisa [1]等提出的自动水体提取指数（Automated Water Extraction Index, AWEI）。
Ji等[26]的研究发现,MNDWI在勾画水体时比NDWI效果好,而且阈值也最稳定。对于Landsat 8 OLI数据,MNDWI公式表达如下：
MNDWI = ( B 3 - B 6 ) ( B 3 + B 6 ) （1）
式中：B3和B6分别为第3和6波段的反射光谱值[25]。
2014年,Feyisa等[1]提出了一种基于TM多波段的自动水体提取指数（Automated Water Extraction Index, AWEI）,并证实AWEI在提取水体信息方面比MNDWI具有更高的精度。
AWEI是基于TM影像数据提出的,主要目标是通过波段间相减、相加以及给波段赋予不同的系数来达到将水体和非水体像元进行最大程度分离的目的,由两个独立的方程构成,移植到Landsat 8 OLI数据应用时,计算公式为：
AWE I nsh = 4 × ( ρ band 3 - ρ band 6 ) ( 0.25 × ρ band 5 + 2.75 × ρ band 7 ) （2）
AWE I sh = ρ band 2 + 2.5 × ρ band 3 - 1.5 × ( ρ band 5 + ρ band 6 ) - 0.25 × ρ band 7 （3）
式中：ρ是Landsat 5 TM光谱波段的反射值,band 2（蓝色）、band 3 （绿色）、band 5 （NIR）、 band 6 （短波红外）、band 7 （短波红外）。
AWEInsh主要用来去除容易与水体混淆的黑色建筑地表,而AWEIsh主要是通过移除AWEInsh中不能有效去除的阴影像元而进一步提升水体提取的精度。 AWEInsh的下标“nsh”表示该指数适合于阴影不是主要问题的地方;AWEIsh的下标“sh”表示该指数主要用来移除阴影,从而提升阴影区域或者其他黑色地表的水体提取精度。
饶萍等[27]在研究建设用地信息提取时,创建了一种三波段水体指数（TB）归一化三波段指数（Normalized Difference Three Bands Index,NDTBI）用于建设用地信息的提取,对于Landsat 8数据,其公 式为：
NDTBI = ( B 7 + B 6 - B 4 ) ( B 7 + B 6 + B 4 ) （4）
式中：B7、B6、B4分别为第7、6和4波段的反射光谱值。
本文研究发现,NDTBI具有压制植被信息而突出水体和低亮度的建设用地信息的作用,从而NDTBI能够弥补MNDWI和AWEI水体指数的不足,而成为本文中的“水体提取辅助指数”。

4 结果对比分析

4.1 水体指数阈值分析

对研究区的Landsat 8 OLI图像应用MNDWI,AWEI和NDTBI3种指数进行阈值分析,发现在研究区全域范围内,没有一种指数能够将水体信息与背景精确地分离开来,如图3-5所示。
将图3与图2中真实水体相比较,对于AWEI,AWEInsh的最佳阈值分割点为-0.375时,研究区内较大的线状河流的分离效果很好（如测试点1和2）,但细小河流信息的分离效果较差（测试点3）,并且混入大量阴影及少量的暗色建筑地物;当AWEIsh阈值分割点为-0.046时,面状水体以及山体阴影的分离效果均较好,但河流信息提取效果则较差;当AWEIsh的阈值分割点为-0.202时,虽然能提取到更多的线状河流（测试点1和2）,但仍然未能提取到细小河流,且此时低反射的暗色地表与部分河流之间连成一片,分离性较差;当AWEIsh的阈值分割点为-0.358时,细小的河流被提取出来,但大量阴影和建设用地也被提取出来。
对于MNDWI,在3个阈值分割点-0.002、-0.120和-0.234处分别与AWEIsh3个阈值点-0.046、-0.202和-0.358的分离效果相似。
而对于NDTBI,在3个阈值点0.282、0.396和0.508处具有和AWEIsh及MNDWI相近的分离效果,但在对应的阈值处,线状河流区域与多提出的非水体区域的分离性要好于前二者。
通过对3个指数的阈值图像进行比较分析,发现：面状水体信息与背景的分离在MNDWI阈值为-0.002处效果最好,大的线状河流与背景的分离在AWEInsh为-0.375且AWEIsh为-0.202时效果最好,受底物影响较大的细小河流与背景的分离在NDTBI阈值为-0.508处效果最好。

4.2 基于指数阈值的分类分区

通过2.3中的水体指数阈值结果显示,无论使用任何一种指数,仅仅使用阈值法均不能完全剔除阴影,暗色地物等干扰因素的影响,但在特定的区域范围内,最佳指数、最佳阈值的选择与水体的类型（如河流、水库、湖泊等）、大小以及干扰因素等具有很大的相关性。
鉴于此,本文利用指数阈值图像对水体进行分类分区,基本思路是：首先,根据某一指数最佳阈值点,对研究区的水体进行分类,分为面状水体与线状水体（溪流）,线状水体又分为受暗色地物影响,受阴影影响和受底物影响3种类型。
在分类的基础上,按照不同类别用创建缓冲区的方式进行二值分区。首先将整个研究区按照水体形状进行分区,分为河流缓冲区域（River buffer）与非两个区域;在河流缓冲区内按照受底物影响的大小,分为受底物影响大的缓冲区域（Small river buffer）和非两个区域;在非河流缓冲区域内,按照受阴影、暗色地物的影像情况分为非水体的缓冲区域（Nonwater buffe）和水体所在区域。
缓冲区的创建方法：（1）河流缓冲区的创建,在最佳指数阈值点分割图像上,创建河流缓冲区（River buffer）,由于影响水体提取精度的主要因素是暗色建筑地表和山体阴影,因此创建河流缓冲区时须遵循如下的原则：包含河流但不包含建设用地和阴影的情况下,缓冲区的大小和形状无需严格要求[27]。受底物影响大的河流缓冲区（Small river buffer）是河流缓冲区一部分,其包含的是指数图像中的细小河流部分。（2）非实体缓冲区的创建,只要将最佳指数阈值点图像中的响应类的非水体信息包含即可。
缓冲区图像均为二值图像,为后续水体信息的提取提供分区分层的依据。

4.3 指数阈值分区分层提取法

综合3.3节和3.4节,对3个指数的阈值分析、对水体与背景的分离效果分析及对不同水体类型进行分区的基础上,分别应用MNDWI、AWEInsh、NDTBI来构建基于单个指数和分类分区的决策树（如图6 中的DTMNDWI、DTAWEI和DTNDTBI）,同时,根据3种指数对不同类型水体的分离效果,按照分类别应用最佳指数最佳阈值进行分层分类的方法,构建综合决策树（如图6中的DTUI）。
Fig. 6 Decision classifier trees based on the optimal partition and indexes

图6 基于最优分区的指数决策树分类法

注：DTMNDWI、DTAWEI、DTNDTBI 分别为MNDWI、AWEI、NDTBI单一指数与空间信息联合的分层分类树,DTUI为3个指数与空间信息联合的分层分类树;W代表水体, NW代表非水体

执行图6中的4个决策树DTMNDWI、DTAWEI、DTNDTBI和DTUI,并利用前期解译得到的真实水体地图（图2）进行精度评价,评价结果如表1所示。
Tab. 1 Summary of accuracy assessment at the two test sites

表1 基于单一指数阈值分区的分层提取法精度评价

方法 测试点 生产者精度 用户精度 误分率 漏分率 总误差 Kappa系数
DTMNDWI 1 91.57 80.31 19.69 8.43 28.12 0.8512
2 90.70 86.12 13.88 9.30 23.18 0.8818
3 84.11 82.30 17.70 15.89 33.59 0.8288
DTAWI 1 91.03 89.36 10.64 8.97 19.61 0.8990
2 81.73 91.28 8.72 18.27 26.99 0.8606
3 80.93 85.53 14.47 19.07 33.54 0.8286
DTNDTBI 1 85.25 93.09 6.91 14.75 21.66 0.8869
2 74.92 96.57 3.43 25.08 28.51 0.8419
3 86.06 85.85 14.15 13.94 28.09 0.8570
DTUI 1 92.30 87.79 12.21 7.70 19.91 0.8968
2 90.70 85.58 14.42 9.30 23.72 0.8789
3 88.26 85.95 14.05 11.74 25.79 0.8685
因Kappa系数是所有精度指标中最全面利用混淆矩阵信息的一个综合分类精度指标,故本文采用Kappa系数来进行分析。
在测试点1、2和3处,方法DTMNDWI的Kappa系数分别为0.8512、0.8818和0.8288,3个测试点的Kappa系数平均值为0.863;方法DTAWEI的Kappa系数分别为0.899、0.8606和 0.8286,3个测试点Kappa系数的平均值为0.854;方法DTNDTBI的Kappa系数分别为0.8869、0.8419和 0.857,3个测试点Kappa系数的平均值为0.862;方法DTUI的Kappa系数分别为0.8968,0.8789,0.8685,3个测试点Kappa系数的平均值为0.881。
根据上述分析,本文采用基于3种指数阈值分区的水体信息分层方法提取研究区的水体信息,结果如图7所示。
Fig. 7 Water extraction results based on the united method of the optimal partition and the optimal indexes at the three test sites

图7 基于最优分区和最优指数联合的水体信息提取结果

5 结论

本文以决策分类树作为框架模型,通过对改进的归一化水体指数（MNDWI）、自动水体提取指数（AWEI）和归一化三波段指数（NDTBI）3种指数的阈值图像分析,对水体进行分类分区,构建以MNDWI、AWEI、NDTBI单一决策树和3个指数联合的决策树进行水体信息的提取,并对3个测试点分别进行了精度评价,其平均Kappa系数分别为0.863、0.854、0.862和0.881。结果表明：基于3个指数构成的最优分区和最优指数联合决策树的水体信息提取方法精度最高,说明采用该法用于水体信息提取可得到最佳效果。
将研究区按照指数进行分类分区,在小区域内实现不同水体类型的分层提取,这是本文提高精度的关键环节。在分类分区的基础上,无论使用单指数方法还是混合指数方法均能满足精度要求,但相比较而言,面状水体的提取选用MNDWI效果最好;线状水体的提取,选用AWEI效果较好;当水体河道很宽,河流水面小,受水体地面的砂石影响较大时,选用NDTBI效果较好;而利用3种指数在不同类型水体提取上的优势,将其结合起来构建最优分区和最优指数联合的决策树提取方法具有最佳的提取精度。

The authors have declared that no competing interests exist.

参考文献

原文顺序 | 文献年度倒序 | 文中引用次数倒序
[1]
Feyisa G L, Meilby H, Fensholt R, et al.Automated water extraction index: A new technique for surface water mapping using Landsat imagery[J]. Remote Sensing of Environment, 2014,140(1):23-35.Classifying surface cover types and analyzing changes are among the most common applications of remote sensing. One of the most basic classification tasks is to distinguish water bodies from dry land surfaces. Landsat imagery is among the most widely used sources of data in remote sensing of water resources; and although several techniques of surface water extraction using Landsat data are described in the literature, their application is constrained by low accuracy in various situations. Besides, with the use of techniques such as single band thresholding and two-band indices, identifying an appropriate threshold yielding the highest possible accuracy is a challenging and time consuming task, as threshold values vary with location and time of image acquisition. The purpose of this study was therefore to devise an index that consistently improves water extraction accuracy in the presence of various sorts of environmental noise and at the same time offers a stable threshold value. Thus we introduced a new Automated Water Extraction Index (AWEI) improving classification accuracy in areas that include shadow and dark surfaces that other classification methods often fail to classify correctly. We tested the accuracy and robustness of the new method using Landsat 5 TM images of several water bodies in Denmark, Switzerland, Ethiopia, South Africa and New Zealand. Kappa coefficient, omission and commission errors were calculated to evaluate accuracies. The performance of the classifier was compared with that of the Modified Normalized Difference Water Index (MNDWI) and Maximum Likelihood (ML) classifiers. In four out of five test sites, classification accuracy of AWEI was significantly higher than that of MNDWI and ML ( < 0.01). AWEI improved accuracy by lessening commission and omission errors by 50% compared to those resulting from MNDWI and about 25% compared to ML classifiers. Besides, the new method was shown to have a fairly stable optimal threshold value. Therefore, AWEI can be used for extracting water with high accuracy, especially in mountainous areas where deep shadow caused by the terrain is an important source of classification error.

DOI

[2]
Alderman K, Turner L R, Tong S L.Floods and human health: A systematic review[J]. Environment International, 2012,47:37-47.Floods are the most common type of disaster globally, responsible for almost 53,000 deaths in the last decade alone (23:1 low- versus high-income countries). This review assessed recent epidemiological evidence on the impacts of floods on human health. Published articles (2004 2011) on the quantitative relationship between floods and health were systematically reviewed. 35 relevant epidemiological studies were identified. Health outcomes were categorized into short- and long-term and were found to depend on the flood characteristics and people's vulnerability. It was found that long-term health effects are currently not well understood. Mortality rates were found to increase by up to 50% in the first year post-flood. After floods, it was found there is an increased risk of disease outbreaks such as hepatitis E, gastrointestinal disease and leptospirosis, particularly in areas with poor hygiene and displaced populations. Psychological distress in survivors (prevalence 8.6% to 53% two years post-flood) can also exacerbate their physical illness. There is a need for effective policies to reduce and prevent flood-related morbidity and mortality. Such steps are contingent upon the improved understanding of potential health impacts of floods. Global trends in urbanization, burden of disease, malnutrition and maternal and child health must be better reflected in flood preparedness and mitigation programs.

DOI PMID

[3]
Bond N R, Lake P S, Arthington A H.The impacts of drought on freshwater ecosystems: An Australian perspective[J]. Hydrobiologia, 2008,600:3-16.

[4]
Charron D F, Thomas M K, Waltner-Toews D, Aramini J J, Edge T, Kent R A, et al.Vulnerability of waterborne diseases to climate change in Canada: A review[J]. Journal of Toxicology and Environmental Health-Part a-Current Issues, 2004,67: 1667-1677.This project addresses two important issues relevant to the health of Canadians: the risk of waterborne illness and the health impacts of global climate change. The Canadian health burden from waterborne illness is unknown, although it presumably accounts for a significant proportion of enteric illness. Recently, large outbreaks with severe consequences produced by E. coli O157:H7 and Cryptosporidium have alarmed Canadians and brought demands for political action. A concurrent need to understand the health impacts of global climate changes and to develop strategies to prevent or prepare for these has also been recognized. There is mounting evidence that weather is often a factor in triggering waterborne disease outbreaks. A recent study of precipitation and waterborne illness in the United States found that more than half the waterborne disease outbreaks in the United States during the last half century followed a period of extreme rainfall. Projections of international global climate change scenarios suggest that, under conditions of global warming most of Canada may expect longer summers, milder winters, increased summer drought, and more extreme precipitation. Excess precipitation, floods, high temperatures, and drought could affect the risk of waterborne illness in Canada. The existing scientific information regarding most weather-related adverse health impacts and on the impacts of global climate change on health in Canada is insufficient for informed decision making. The results of this project address this need through the investigation of the complex systemic interrelationships between disease incidence, weather parameters, and water quality and quantity, and by projecting the potential impact of global climate change on those relationships.

DOI PMID

[5]
Kondo H, Seo N, Yasuda T, Hasizume M, Koido Y, Ninomiya N, et al.Post-flood-infectious diseases in Mozambique[J]. Prehospital and Disaster Medicine, 2002,17:126-133.

[6]
Lake P S.Ecological effects of perturbation by drought in flowing waters[J]. Freshwater Biology, 2003,48:1161-1172.

[7]
Li K Z, Wu S H, Dai E F, Xu Z C.Flood loss analysis and quantitative risk assessment in China[J]. Natural Hazards, 2012,63:737-760.

[8]
Lira J.Segmentation and morphology of open water bodies from multispectral images[J]. International Journal of Remote Sensing, 2006,27:4015-4038.A number of problems in remote sensing require the segmentation of specific spectral classes such as water bodies, clouds or forested areas. Further analysis of these classes may include the calculation of optical reflectance values such as chlorophyll concentration, absolute reflectivity or vegetation indices. To derive reliable measurements of these variables, a precise segmentation from the rest of the image of the spectral classes is needed. In this work, we propose a new methodology to segment open water bodies based on a variant of principal component analysis (PCA). In this variant, information about the spectral class model of the water bodies to be separated in the feature space is required. This information is input by means of a training field encircling a set of pixels representative of this spectral model. A training field for land cover is also defined. This PCA variant produces two sets of multispectral bands, one for water bodies and one for land cover types. The first two bands of each set are input into a fuzzy clustering procedure. By using a merging process, the clusters are merged into two classes: water bodies and the rest of the image. From this, a logic bitmap image is obtained. The pixels of the bitmap consist of ON for water bodies and OFF for the rest of the image. The bitmap is then used to obtain morphological parameters of the water bodies.

DOI

[9]
Sethre P R, Rundquist B C, Todhunter P E.Remote detection of prairie pothole ponds in the Devils Lake Basin, North Dakota[J]. GIS science and Remote Sensing, 2005,42:277-296.A decadal-scale wet spell in the closed Devils Lake basin of North Dakota has resulted in increases in the elevation and extent of the basin's terminal lakes—Devils and Stump—as well as increases in the size and number of small prairie pothole ponds. Changes in lake surface area have been studied thoroughly, whereas the fluctuations in pond surface area have been virtually ignored. We use a subpixel classification technique in combination with a Landsat TM and ETM+ Band 5 (middle infrared; 1,550-1,750 nm) density slice to improve estimates of changes in the combined area of ponds in the basin for selected years between 1991 and 2002. The resulting information is a first step toward more accurate assessment of the impact of wetland flooding on the region.

DOI

[10]
Jain S K, Singh R D, Jain M K, et al.Delineation of flood-prone areas using remote sensing technique[J]. Water Resources Management, 2005,19:337-347.Flood problems resulting due to heavy rainfall and drainage congestion are being regularly experienced in plain areas of Bihar, India. Due to this problem, power plant located in Koa catchment, Kahalgaon, Bihar, is faces huge loss at the time of flood. In this paper, the flood-affected areas in Koa catchment have been mapped using remote sensing satellite data (IRS LISS III, 1999 and Landsat TM, 1995). A range of image processing techniques has been used, including simple density slicing, Tasseled Cap Transformation and water-specific index. The results obtained using different approaches have been analysed. Result indicates that a Normalized Difference Water Index (NDWI) based approach produced best results. Copyright Springer Science + Business Media, Inc. 2005

DOI

[11]
Jain S K, Saraf A K, Goswami A, et al.Flood inundation mapping using NOAA AVHRR data[J]. Water Resources Management, 2006,20:949-959.The river Brahmputra flowing through the state of Assam (India) floods every year. The analysis of spatial extent and temporal pattern of flood-inundated areas is of prime importance for mitigation of floods. With the development of remote sensing techniques, flood mapping for large areas can be done easily. In case of flood affected area mapping of large area it will not be feasible to use high-resolution data, because the whole area will be covered in number of scenes. Therefore use of NOAA (National Oceaongraphic Atmospheric Administrative) data is quite useful in such type of studies. NOAA-AVHRR (Advanced Very High Resolution Radiometer) data have the potential for flood monitoring due to high frequency of global coverage, wide swath, high repetivity and low cost. In this study, NOAA-AVHRR data have been used for mapping of flood-affected area during the year 2003. On the basis of spectral characteristics of land and water, a methodology for water identification is presented. The maximum spatial extent of floods, generated by compiling the available cloud free maps, is informative about flood damages. Analysis of results reveals that in the months of July and August almost 25-30% of the area was flood affected. Also the result indicates that in some districts, the flood-affected area is very high. Copyright Springer Science + Business Media, Inc. 2006

DOI

[12]
McFeeters S K. The use of Normalized Difference Water Index (NDWI) in the delineation of open water features[J]. International Journal of Remote Sensing, 1996,17:1425-1432.ABSTRACT The Normalized Difference Water Index (NDWI) is a new method that has been developed to delineate open water features and enhance their presence in remotely-sensed digital imagery. The NDWI makes use of reflected near-infrared radiation and visible green light to enhance the presence of such features while eliminating the presence of soil and terrestrial vegetation features. It is suggested that the NDWI may also provide researchers with turbidity estimations of water bodies using remotely-sensed digital data.

DOI

[13]
Rogers A S, Kearney M S.Reducing signature variability in unmixing coastal marsh Thematic Mapper scenes using spectral indices[J]. International Journal of Remote Sensing, 2004,25:2317-2335.Evidence of the rapid losses of coastal marshes calls for the application of remote sensing data. Nonetheless, many features indicative of incipient marsh loss, such as widening of tidal creeks and formation of small ponds, are often not readily detectable at the nominal 30 m 30 m resolution of Thematic Mapper (TM) imagery, the general source for conventional satellite sensor-based data on wetlands. Spectral mixture modelling, where the proportional representation of land cover types can be estimated within pixels, offers a potential solution to the problem of assessing initial indications of loss in coastal marshes. Nevertheless, the simple linear mixture models most commonly employed can be subject to significant errors when applied to marshes due to the considerable variety of soil/sediment types in these environments. A new method is presented here which not only successfully reduces the spectral variability of soils, but also in the other principal components of vegetation and water, to a three-endmember model for estimating their fractional representation in TM data. Fractional representations for these variables in this method yield more reliable results than those obtained from unmixing corrected reflectance data. Though unmixing individual scenes may still be best achieved by extensive ground-referencing and image analysis, this technique is a robust approach for large-scale, semi-automated processing of many scenes in investigations of marsh surface condition.

DOI

[14]
Xu H Q.Modification of normalised difference water index (NDWI) to enhance open water features in remotely sensed imagery[J]. International Journal of Remote Sensing, 2006,27:3025-3033.The normalized difference water index (NDWI) of McFeeters (1996) was modified by substitution of a middle infrared band such as Landsat TM band 5 for the near infrared band used in the NDWI. The modified NDWI (MNDWI) can enhance open water features while efficiently suppressing and even removing built-up land noise as well as vegetation and soil noise. The enhanced water information using the NDWI is often mixed with built-up land noise and the area of extracted water is thus overestimated. Accordingly, the MNDWI is more suitable for enhancing and extracting water information for a water region with a background dominated by built-up land areas because of its advantage in reducing and even removing built-up land noise over the NDWI.

DOI

[15]
沈占锋,夏列钢,李均力,等.采用高斯归一化水体指数实现遥感影像河流的精确提取[J].中国图象图形学报,2013,18(4):421-428.河流精确提取在水资源调查、利用、变化检测及大型水利设施建设评估等方面具有非常重要的意义。通常的河流信息提取方法受影像中云、冰雪、山体阴影、大型湖泊等的干扰较大,大范围的适用性有限。以Landsat卫星遥感数据为数据源,在归一化差异水指数(NDWI)计算的基础上,首次提出采用高斯归一化水体指数(GNDWI)提取河流水体的模型,使得指数能够更大程度上保证河流提取的连续性,并通过DEM的辅助实现了其他干扰信息的去除。通过对伊犁河试验区河流信息提取的实验结果表明,该方法除了能够实现对复杂多样的河流水体信息进行自动提取外,还可有效去除阴影等信息的混淆,并能够达到较高的河流提取精度。

DOI

[Shen Z F, Xia L G, Li J L, et al.Automatic and high—precision extraction of rivers from remotely sensed images with Gaussian normalized water index[J]. Journal of Image and Graphics, 2013,18(4):421-428.

[16]
杨树文,薛重生,刘涛,等.一种利用TM影像自动提取细小水体的方法[J].测绘学报,2010,39(6):611-617.提出一种利用TM影像自动提取山区细小水体的多波段谱间关系改进方法。该方法在典型谱间关系法的基础上，针对水体与阴影在蓝绿光波段亮度值降低速率差异较大的特征，基于差值运算，构建了新的多波段谱间关系水体提取模型。首先利用该模型将水体从其他地物及阴影中分离出来，然后基于数学形态学膨胀滤波算法进行空洞填充和短线连接，最后通过图像细化算法实现目标的细化。经过实验比较表明，该方法克服了许多水体提取模型只能有效提取较大面积水体的缺点，除了能够对山区的细小水体进行高精度自动提取外，还能够有效地去除阴影等干扰信息。

DOI

[Yang S W, Xue C S, Liu T, et al.A method of small water information automatic extraction from TM remote sensing images[J]. Acta Geodaetica of Cartographica Sinica, 2010,39(6):611-617. ]

[17]
骆剑承,盛永伟,沈占锋,等.分步迭代的多光谱遥感水体信息高精度自动提取[J].遥感学报,2009,13(4):610-615.以LANDSAT卫星遥感数据为信息源, 在归一化差异水指数(NDWI)计算的基础上, 采用“全域―局部”的分步迭代空间尺度转换机制, 将全域分割、全域分类、局部分割与分类等计算过程有机地结合起来, 分阶段地融合了水体信息提取所需的不同层次知识, 并建立迭代算法实现了水体最佳边缘的逐步逼近, 获得了高精度的水体信息提取。通过对青藏高原试验区湖泊信息提取的实验表明, 该方法除了能够实现对复杂多样的水体信息进行高精度自动提取外, 还可有效避免与阴影等信息的混淆。

DOI

[Luo J C, Sheng Y W, Shen Z F, et al.Step Iterative mult-ispectral remote sensing information on high-precision automatic extraction of water[J]. Journal of Remote Sensing, 2009,13(4):610-615. ]

[18]
Sun F, Sun W, Chen J, Gong P.Comparison and improvement of methods for identifying waterbodies in remotely sensed imagery[J]. International Journal of Remote Sensing ,2012,33:6854-6875.This article first examines three existing methods of delineating open water features, i.e. the normalized difference water index (NDWI), the modified normalized difference water index (MNDWI) and a method combining the near-infrared (NIR) band and the maximum likelihood classification. We then propose two new methods for the fast extraction of water features in remotely sensed imagery. Our first method is a pixel-based procedure that utilizes indices and band values. Based on their characteristic spectral reflectance curves, waterbodies are grouped into three types clear, green and turbid. We found that the MNDWI is best suited for identifying clear water. Green water has its maximum reflectance in Landsat Thematic Mapper (TM) band 4 (NIR band), whereas turbid water has its maximum reflectance in TM band 5 (mid-infrared band). Our second method integrates our pixel-based classification with object-based image segmentation. Two Landsat scenes in Shaanxi Province, China, were used as the primary data source. Digital elevation models (DEMs) and their derived slope maps were used as ancillary information. To evaluate the performance of the proposed methods, extraction results of the three existing methods and our two new methods were compared and assessed. A manual interpretation was made and used as reference data. Results suggest that our methods, which consider the diversity of waterbodies, achieved better accuracy. Our pixel-based method achieved a producer's accuracy of 92%, user's accuracy of 90% and kappa statistics of 0.91. Our integrated method produced a higher producer's accuracy (95%), but a lower user's accuracy (72%) and kappa statistics (0.72), compared with the pixel-based method. The advantages and limitations of the proposed methods are discussed.

DOI

[19]
Jiang Z, Qi J, Su S, Zhang Z, Wu J.Water body delineation using index composition and HIS transformation[J]. International Journal of Remote Sensing, 2012,33:3402-3421.The use of satellite data for mapping water bodies is important for environmental management. Previous approaches exhibit limited applicability in southeastern China, given its complex and heterogeneous landscapes as well as the difficulty to obtain cloud-free images. To overcome these problems, we proposed an approach using index composition and HIS (hue, intensity and saturation) transformation. First, a colour image was generated using three indices: the Normalized Difference Built-up Index (NDBI), the Normalized Difference Vegetation Index (NDVI) and the Modified Normalized Difference Water Index (MNDWI). Then, HIS transformation was employed to extract water bodies and remove hill shadows. Another colour image, composed by Landsat TM (Thematic Mapper) 4, NDVI, TM1 and HIS transformation, was further applied to separate water bodies from residual shadows. This approach was tested and verified with Landsat TM images in the Tiaoxi watershed, southeastern China. The results indicated the high accuracy and promising applicability of this new approach for fragmented landscapes, given its insensitivity to seasonal and subjective factors.

DOI

[20]
Sheng Y W, Shah C A, Smith L C.Automated image registration for hydrologic change detection in the lake-rich Arctic[J]. IEEE Geoscience and Remote Sensing Letters, 2008,5:414-418.

[21]
Verpoorter C, Kutser T, Tranvik L.Automated mapping of water bodies using Landsat multispectral data[J]. Limnology and Oceanography-Methods, 2012,10:1037-1050.

[22]
Billa L.Modelling rainfall intensity from NOAA AVHRR data for operational flood forecasting in Malaysia[J]. International Journal of Remote Sensing, 2006,27(23):5225-5234.

[23]
闫霈,张友静,张元.利用增强型水体指数(EWI)和GIS去噪音技术提取半干旱地区水系信息的研究[J].遥感信息,2007(6):62-67.准确提取流域水系信息是进行水资源开发利用的首要任务。利用NDWI和MNDWI提取半干旱地区水系，无法有效的区分水系内的半干涸河道与背景噪音。本文在分析半干旱地区水系与背景噪音反射特点的基础上，提出了增强型水体指数EWI（Enhanced Water Index），有效地区分了半干涸河道与背景噪音。在利用形状指数去噪音方法的基础上，使用GIS技术去除背景噪音，弥补了形状指数去噪音方法的缺陷，更好的去除了水系提取过程中混入的背景噪音。综合使用上述两种方法，可以快速、准确、简便地提取出半干旱地区的水系。为此类地区水资源的开发利用打下了基础。

DOI

[Yan P, Zhang Y J, Zhang Y.Study on extracting water system information in semi-arid area using Enhanced Water Index (EWI) and GIS denoising technology[J]. Remote Sensing Information, 2007,6:62-67. ]

[24]
杨莹,阮仁宗.基于TM影像的平原湖泊水体信息提取的研究[J].遥感信息,2010(3):60-64.

[Yang Y, Ruan R Z.Study on extracting water information of plain lakes based on TM images[J]. Remote Sensing Information, 2010,3:60-64. ]

[25]
汪金花,张永彬,孔改红.谱间关系法在水体特征提取中的应用[J].矿山测量,2004,12(4):30-32.

[Wang J H, Zhang Y B, Kong G H.The application of spectrum-photometric method in water extractions[J]. Mine Surveying, 2004,12(4):30-32. ]

[26]
Ji L, Zhang L, Wylie B.Analysis of dynamic thresholds for the normalized difference water index. Photogrammetric Engineering and Remote Sensing, 2009,75:1307-1317.The normalized difference water index (NDWI) has been successfully used to delineate surface water features. However, two major problems have been often encountered: (a) NDWIs calculated from different band combinations [visible, near-infrared, or shortwave-infrared (SWIR)] can generate different results, and (b) NDWI thresholds vary depending on the proportions of subpixel water/non-water comp...

DOI

[27]
饶萍,王建力,王勇.基于多特征决策树的建设用地信息提取[J].农业工程学报,2014,30(12):233-240.城乡交接带的土地利用/覆盖类型兼具城镇和农村的典型特征。为了解决土地覆盖类型复杂、存在"同谱异物"现象的西部山区环境中建设用地信息难以精确提取的问题，该文提出一种包含多个特征节点的决策树分类法，该方法以Landsat-8影像为主要数据源，以决策树分类法为框架，结合地物光谱特征及空间特征，建立以4种归一化指数（归一化三波段指数normalized difference three bands index，NDTBI；归一化建筑指数normalized difference building index，NDBI；改进的归一化水体指数modified normalized difference water index，MNDWI；归一化植被指数normalized difference vegetation index，NDVI）、支持向量机（support vector machine，SVM）分类结果和河流缓冲区作为特征节点的决策树分类器，对贵州省毕节市城乡交接带建设用地专题信息进行提取。NDTBI是该文新构建的指数，取名为归一化三波段指数，目的是为了弥补归一化建筑指数NDBI的不足；支持向量机分类结果的使用在多指数法的基础上提高了地物的可分离性；以构建河流缓冲区的方式加入的地物空间信息, 进一步提高了信息提取的精确性。由于决策树特征节点的构建过程是利用先验知识来优化特征值和提高精度的过程，克服了利用单一指数法、多指数法及单独使用模式识别法中出现的问题，精度评价结果显示总体精度达到了97.52%。为了验证方法的推广性，采用毕节市七星关区中心城区遥感影像数据该方法进行验证，精度评价结果显示总体精度达到98.03%。

DOI

[Rao P, Wang J L, Wang Y.Extraction of construction land based on multi-features decision tree classification[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2014,30(12):233-240. ]

Options
文章导航

/

〈 〉