FY-2C积雪判识方法研究  

Study of Snow Detection Using FY-2C Satellite Data

作　　者：李三妹[1] 闫华[1] 刘诚[1] 

机构地区：[1]国家卫星气象中心,北京100081

出　　处：《遥感学报》2007年第3期406-413,共8页NATIONAL REMOTE SENSING BULLETIN

摘　　要：介绍了利用FY-2C资料进行积雪判识的原理,在阈值法基础上的辅助因子函数积雪判识方法以及相应的FY-2C积雪判识结果精度验证分析等。一般较为常用的卫星遥感积雪判识方法为简单阈值法,由于其带有一定的随机性,很难客观反映下垫面条件差异对阈值选取的影响。以阈值法为基础,将所使用的主要变量以函数形式表达,以海拔高度、地理位置、季节、土地覆盖类型等作为阈值函数的变量,通过大量采样建立起多种阈值函数,从而实现随时空特点变化的阈值实时计算。该方法用于FY-2C积雪判识,较好地解决了FY-2C全圆盘范围内广大区域不同下垫面类型下的实时积雪监测。通过与NOAA-17人机交互积雪判识结果对比分析,该方法的积雪判识精度可达85%左右。Snow/ice, one of the most important earth circles, is very significant to climate research and earth environment study. Snow cover, a basic parameter in snow study, can reflect snow amount most directly. The method of detecting snow cover using satellite data automatically is a very interesting topic for snow study. FY-2C, the third geostationary meteorological satellite launched by China in 2004, has 5 channels in its scanning radiometer, including 3.5-4.0μm, water vapor, visible and split infrared channels, which makes snow detection possible. This paper mainly introduces the principle theory and method of snow detection using FY-2C satellite data. Multi-channel thresholds method is a very common way in snow and cloud detection. However, this method has some randomicity in thresholds choosing properly. To avoid the randomicity caused by multi-channel thresholds method, this paper uses threshold functions to take instead of traditional method for snow detection, which uses altitude, geographical location, season, land cover and so on as the variables. Threshold functions can be established by large amount of sampling to obtain coefficients and their expressions. Compared with NOAA-17 snow detection results derived using interactive method, the precision of FY-2C snow detection using this method can reach up to 85 percents.

关 键 词：FY-2C 积雪覆盖 判识 遥感 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]