摘要
新疆北疆为雪灾多发区,准确地监测积雪分布及雪深具有十分重要的意义,可为雪灾多发区在冬季做好防灾减灾工作提供依据。目前MODIS数据主要运用归一化差分雪盖指数(NDSI)对积雪进行提取。NDSI是基于雪对可见光与近红外波段的反射特性和反射差相对大小的一种测量方法,然而它只能将像元辨别为积雪或非积雪,满足不了高精度的流域制图及雪盖提取要求。以新疆北疆为研究对象,应用MODIS数据和线性光谱混合模型提取像元内积雪所占比例(积雪盖度),并与NDSI建立关系,讨论了NDSI是否可以作为估算积雪盖度的标准,进而提高NDSI在积雪监测中的应用精度。结果表明,二者具有较好的线性关系,经过野外实测点的验证,25个点的平均绝对误差是0.06。此外,以MODIS数据与地面实测雪深资料为基础,通过分析和研究MODIS窗区通道的光谱特点,探讨了适合于积雪雪深监测最佳的MODIS通道,初步建立起应用MODIS资料监测北疆地区积雪深度的反演模型。
The snow disaster often takes place in the north of Xinjiang. So it is of great significance to exactly monitor the snow distribution and snow depth in the northern Xinjiang, which can provide scientific basis for snow disaster prevention and reduction. In recent years, NDSI is mainly used to abstract the snow cover with MODIS data. The NDSI is a spectral ratio that takes advantage of the spectral difference of snow in short-wave infrared and visible spectral bands. It can only discern one pixel into snow or other features, and can not satisfy accurate drainage basin snow cover mapping and snow parameter extracting. In this study, linear spectrum mixing model was used to abstract snow fraction in the north of Xinjiang. Then we established the relationship between snow fraction and NDSI and evaluated whether NDSI can be used to estimate the cover rate of snow within a 250m pixel. The result showed that they had good linear relationship. The mean absolute error for 25 true measured points was 0.06. Moreover, we analyzed the correlation between snow depth and the reflected spectrum of snow and compared the true measured snow reflected spectrum with the image reflected spectrum. The most sensitive bands to snow depth were chosen. At last, the snow depth-inversing model was built.
出处
《自然灾害学报》
CSCD
北大核心
2008年第5期52-57,共6页
Journal of Natural Disasters
基金
国家自然科学基金资助项目(70361001)
关键词
积雪盖度
雪深
混合像元
新疆
snow coverage
snow depth
mixed pixel
Xinjiang
