摘要
暴雨后水体含沙量增加使其光谱反射率发生变化,导致利用现有水体指数方法不能很好地提取水体信息。为有效提取洪水及其他类型水体,该文基于GF-2影像,根据暴雨后水体光谱信息变化规律,提出一种超绿水体指数(Ultra-Green Water Index,UGWI),并通过实验对比分析UGWI与归一化差异水体指数法、单波段阈值法、单波段与阴影水体指数结合法的水体提取精度,结果表明:UGWI能够准确识别不同泥沙含量以及阴影下的水体,总体精度达95.75%,Kappa系数为0.91,效果优于其他几种方法;选取不同时间、不同地区、不同空间分辨率的影像进行水体提取,UGWI结果亦优于其他几种方法,总体精度达92.33%,Kappa系数为0.90;UGWI相比其他对比方法,抗干扰性更强,提取的水体信息更全面,可识别的水体类型更多,对洪区水体提取更具普适性。
The increase of sediment concentration in water body after storm results in the change of water spectral reflectance,which leads to the low accuracy of existing water index methods in remote sensing retrieval of water body.In order to effectively extract flood and other types of water body,based on GF-2 image,according to the change law of water spectral information after storm,an ultra-green water index(UGWI)was proposed,and UGWI was compared with normalized difference water index method,single band threshold method,single band and shadow water index method by water extraction experiments.The experimental results show that UGWI can accurately identify the water body with different sediment concentrations from clear water to flood,as well as the water body under shadow.The overall accuracy is 95.75%,and the Kappa coefficient is 0.91,which is better than other extraction methods of water body.The images of different time,different area and different spatial resolution were selected to extract water information.The results show that UGWI is superior to other methods,the overall accuracy is 92.33%,and Kappa coefficient is 0.90.The study shows that UGWI has stronger anti-interference ability compared with other methods,the extracted water information is more comprehensive,the identifiable water types are more abundant,and it is more suitable for water body extraction in flood areas.
作者
段纪维
钟九生
江丽
钟淼淼
DUAN Ji-wei;ZHONG Jiu-sheng;JIANG Li;ZHONG Miao-miao(School of Geography and Environmental Science,Guizhou Normal University,Guiyang 550025,China)
出处
《地理与地理信息科学》
CSCD
北大核心
2021年第3期35-41,共7页
Geography and Geo-Information Science
基金
国家自然科学基金项目(41661081)
贵州省科技计划项目(黔科合平台人才[2017]5726-56)。
关键词
GF-2影像
水体指数
波段阈值
水体提取
雨后洪区
GF-2 image
water body index
band threshold
extraction of water body
flood area after rain
