基于OLI影像和DEM的山区水体提取方法  

Method of Water-body Extraction in Mountainous Area Based on OLI Images and DEM

作　　者：赵冰雪 王雷[2] 胡和兵[2] ZHAO Bingxue;WANG Lei;HU Hebing(School of Geographic and Oceanographic Sciences,Nanjing University,Nanjing 210023,China;College of Resources and Environment,Chizhou University,Chizhou 247000,China)

机构地区：[1]南京大学地理与海洋科学学院,江苏南京210023 [2]池州学院资源环境学院,安徽池州247000

出　　处：《水文》2019年第4期34-39,83,共7页Journal of China Hydrology

基　　金：安徽省自然科学研究项目(KJ2019A0866;KJ2018A0580);池州学院质量工程项目(2017XKFKC06)

摘　　要：以皖南山区秋浦河流域为研究对象,综合利用OLI影像和DEM数据,针对山区水体提取中出现的阴影问题和河流断线问题提出一种新的解决思路。首先通过比较单波段阈值法、NDWI水体指数法、谱间关系法以及SWI阴影水体指数几种水体提取方法,显示SWI能够剔除大部分山体阴影;其次使用基于数学形态的膨胀滤波和Pavlidis图像细化算法对SWI提取结果进行处理,较短水体的断线得到连接;最后对于水体断线严重的上游河流,利用DEM进行河网的提取,并以提取的河网代替水体指数的提取结果。结果表明,基于OLI和DEM数据进行山区水体的提取,方法简单,提取的面状水体齐全,线状水体连续,精度达到83.2%,该方法可为其他山区地形水体提取提供一定的借鉴。This paper took the Qiupuhe basin in the southern Anhui as research object,and proposed a new solution on the shadow and break line problems of water body extraction in the mountainous area based on OLI images and DEM data.Firstly,SWI can eliminate most of mountain shadows by comparing several methods of single-band threshold,water index of NDWI,spectrum-photometric and SWI(shadow water index);Secondly,the dilation filtering based on mathematical form and image thinning method of Pavlidis were applied to deal with the results of SWI extraction,and the shorter broken line got connection;Finally,for the serious area of water-body broken line in the upstream rivers,DEM was adapted to extract the river network to replace the extraction result of water index.The results show that the new solution is simple,complete on the planar water extraction,continuous in the linear water and high accuracy of 83%.And it can provide reference of water-body extraction for other mountainous terrain.

关 键 词：细小水体 SWI 膨胀滤波 图像细化 DEM 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]