MNDWI和NDWI指数提取富营养化湖泊水体边界的对比研究

Abstract

本研究基于Landsat 8 OLI卫星影像，采用云南省中部地区的5个湖泊，即滇池、抚仙湖、阳宗海、星云湖和杞麓湖作为研究对象，其中滇池、星云湖、杞麓湖是富营养化湖泊，比较了两种水体指数，NDWI和MNDWI提取湖泊水体边界的精确度。结果表明：由于近红外波段能够较好的区分植被和水体，但也带来了藻类覆盖的水域容易被误识别为植被的情况。因此，采用NDWI提取的富营养化湖泊水体面积比实际值偏小，中红外波段对水分含量较为敏感，在富营养化湖泊中采用该指数可以有效的区分陆地和水体，排除藻类的干扰。因此，MNDWI更适合用于富营养化湖泊的水体边界提取。 This study is based on the Landsat 8 OLI satellite images, and uses 5 lakes as sample, which are Dianchi Lake, Fuxian Lake, Yangzonghai Lake, Xingyun Lake and Qilu Lake, there are all located in the central region of Yunnan province, and Dianchi Lake, Xingyun Lake, Qilu Lake are eutrophic lakes. The accuracy of NDWI and MNDWI index for lake water boundary extraction was compared. The results show that the near infrared band can distinguish vegetation and water body better. But it has also brought the algae covered waters that can be mistakenly identified as vegetation. Therefore, the water area of eutrophic lake extracted by NDWI is smaller than the actual value. The infrared band is more sensitive to moisture content. Using this index in eutrophic lakes can effectively distinguish between land and water, eliminate algal interference. Therefore, the MNDWI index is more suitable for water boundary extraction in eutrophic lakes.