基于FY-3A/MERSI中国区域地表类型识别研究

Research on Identification of Land Surface Type in China Based on FY-3A /MERSI

基于风云三号卫星(FY3)搭载的中分辨率光谱成像仪(MERSI)可见光、近红外和热红外通道资料,利用支持向量机分类方法,对整个中国区域4个时段(2012年2月、2012年4月、2012年8月和2012年11月)地表类型进行识别,并进行分类精度验证。结果表明:4个时段分类结果的总体精度都保持在95%以上。可见,基于FY-3A/MERSI数据,利用支持向量机方法得到的分类结果精度非常高,符合实际情况,表明MERSI资料及本文研究方法在地表类型识别上具有其有效性。

This study is based on the visible,near infrared and thermal infrared channel data from the secondgeneration polar-orbiting meteorological satellite of China （FY-3） equipped with the Medium Resolution Spectral Imager （MERSI）.It firstly identified land surface of the entire Chinese region in four times（2012.2,2012.4,2012.8 and 2012.11） by using support vector machine and verified classification accuracy.The results showed that：all the overall accuracy of classification in four times are maintained above 95 ％.It proved that study based FY-3A/MERSI data using support vector machine is highly accuracy and reasonable,and can be validly used in land surface type identification.