Snow Information Abstraction Based on Remote Sensing Data： Taking the North of Xinjiang for Example

This paper proposes an applicable approach for snow information abstraction in northern Xinjiang Basin using MODIS data. Linear spectral mixture analysis （LSMA） was used to calculate snow cover fractions （SF） within a pixel, which was used to establish a regression function with NDSI. In addition, 80 snow depths samples were collected in the study region. The correlation between image spectra reflectance and snow depth as well as the comparison between measured snow spectra and image spectra was analyzed. An algorithm was developed for snow depth inversion on the basis of the correlation between snow depth and snow spectra in the region. The results indicated that the model of SF had a high accuracy with the mean absolute error 0.06 tested by 26 true measured values and the validation for snow depth model using another dataset with 50 sampling sites showed an RMSE of 1.63. Our study showed that MODIS data provide an alternative method for snow information abstraction through development of algorithms suitable for local application.

这篇论文用 MODIS 数据在北 Xinjiang 盆为雪信息抽象建议一条适用的途径。线性光谱混合分析(LSMA ) 被用来在一个象素以内计算积雪部分(SF ) ，它被用来与 NDSI 建立一个回归函数。另外， 80 件雪深样品在学习区域被收集。在象在测量的雪系列和图象系列之间的比较一样的图象系列反射和雪深之间的关联被分析。一个算法在这个区域根据在雪深和雪系列之间的关联为雪深倒置被开发。结果显示 SF 的模型与 0.06 与地点显示出的 50 采样用另一数据集为雪深模型由 26 真观测值和确认测试了的吝啬的绝对误差有高精确性 1.63 的 RMSE。我们的学习证明 MODIS 数据通过对本地申请合适的算法的开发为雪信息抽象提供一个其他的方法。