利用HJ-1卫星数据监测山区水库资源

Mountain Reservoirs Resources Monitoring Based on HJ-1 Satellite Data

遥感技术能够快速、宏观地获得研究区域的数据,已成为湖泊环境动态变化监测的重要技术手段。但是对于山区的湖泊和河流而言,由于沟壑众多,河道狭窄,水体像元多为混合像元,利用现有方法提取水体遥感影像难度较大。水体与植被、城市和土壤等地物在不同波段的光谱反射率的差异是利用遥感手段提取水体信息的基本原理。本文在水体光谱特征分析的基础上,描述了HJ-1卫星数据特征及所采用的山区水体信息的遥感提取方法,并以南湾水库为例,描述了利用植被指数(NDVI)方法识别水体信息的技巧:首先将HJ-1数据的第3和第4波段经过波段运算得到NDVI图像;再将两种遥感图像中NDVI值小于0的像元判识为水体,NDVI值大于0的判识为水库周围的农田,经过计算像元数量得到水体面积信息;最后在各种专业软件的支持下,结合非监督分类的水体识别效果,将水体与其他地物类型区分开来。结果表明,利用HJ-1卫星数据可有效排除其他地物的干扰,显著提高水体监测的精度。

Remote sensing technique has recently become one of the important means for observing and monitoring dynamic changes of large lakes, since it holds the advantage of collecting large-scale geographic information in a convenient and fast way. In the mountain area there are many gullies and the rivers usually are very narrow, so it is too hard to extract water body information with existing method since water body pixels are mixed pixels. The difference of different band reflectance among water, vegetation, urban, soil and other objects is the foundation of extracting water body information by means of the remote sensing data. After analyzing spectral characteristics of the water body information in the remote sense images, in this study we introduced the characteristics of HJ-1 data and the method of extracting water infor- mation in the mountain area by spectral characteristics and NDVI. The NDVI images by case of Nanwan reservoir were extracted from the HJ-1 images of band 3, 4,the pixels with NDVI value less than zero were considered to be water body in the NDVI images and the pixels whose value were higher than zero are considered to the farmland around Nanwan reservoir. The water body information could be extracted by calculating the number of pixels which were considered to water body. And then described effective- ness of extracting water body by specialized software and unsupervised classification by case of Nanwan reservoir in detail. The results showed that HJ-1 data could reduce substantially the influence of other dif- ferent ground objects and extract water body information with a high precision.