湖泊重要入湖河流稳定水像元提取技术  

The Extraction Technology of Stable Water Pixels for the Important Influent Rivers of Lakes

作　　者：吕美婷[1] 赵德华[1] 杨棠武[1] 安树青[1] 

机构地区：[1]南京大学生命科学院,江苏南京210093

出　　处：《环境监控与预警》2013年第5期6-9,共4页Environmental Monitoring and Forewarning

基　　金：江苏省环境监测科研基金(1014)

摘　　要：利用2010年ALOS影像研究了归一化水体指数（NDWI）在两种极端情况下，即像元四边与河岸线最小夹角分别为0°和45°，与离岸距离之间的关系，提出了获得稳定水像元的最小河宽，并以太湖6条重要入湖河流（西苕溪、望虞河、太浦河、乌溪港、梁溪一东清河、武进港）为例，研究了稳定水像元的分布。结果表明，NDWI与离岸距离之间的关系可用Logistic方程进行拟合，两种极端情况的P值均小于0．01；最小夹角为0°和45°的情况下，获得稳定水像元的最小河宽分别为96m和88m，即当河宽大于96m时，可在ALOS影像上获得稳定的水像元；6条人湖河流可获得稳定水像元的长度占总长的4．59％-100％，难以获取稳定水像元的最长距离为0～18．55km。The objectives was to identify the relationship between normalized water index（NDWI） of fiver pixels and their dis tance to river bank using ALOS image in 2010 （ in cases that the smallest angles of pixel sides and fiver bank line were 0° and 45°, respectively） , and determine the least river width for the rivers including stable water pixels, and also quantify the distribution of the stable water pixels in six key rivers, i.e. Xitiaoxi, Wangyu River, Taipu River, Wuxigang, Liangxi-Dongqing River and Wujin- gang. Results suggested that the relation between NDWI and the distance to river bank could be well modeled by Logistic equation, with p-values lower than 0.01 and the thresholds of river width were 96 m and 88 m at both of the two extreme cases （ i. e. , the two cases with the smallest angles of pixel sides and river bank line were 0° and 45°, respectively）. In other word, there are stable water pixels for the river wider than 96 m. For the six rivers, we could obtain stable water pixels in 4.59% - 100% cross-section, and the longest distance of the continual cross-sections in which we couldn＇t obtain stable water oixels were between 0 - 18.55 km.

关 键 词：遥感 太湖 水质 河流 水像元 

分 类 号：X87[环境科学与工程—环境工程]