全球变化与地表参数的定量遥感

GLOBAL CHANGE STUDY AND QUANTITATIVE REMOTE SENSING FOR LAND SURFACE PARAMETERS

从全球变化研究对遥感获取地表时空多变要素信息的需求 ,论述定量遥感中对地表多尺度空间信息作地学描述的新思路和新方法。在分析了变差图描述方法的特点及其不足的基础上 ,提出了直方变差图的概念 ,通过定义驻点和边界点 ,尝试采用直方变差图分解的方法分析地物空间分布规律 ,并通过一个实验对此方法进行了说明。进而基于对地表空间分布信息的提取 ,将可见光、近红外波段遥感信息用于热红外遥感的地温反演 ,利用地表空间结构和大气温度空间分布模式的差异 ,将热红外遥感图像分为地表贡献和大气贡献两部分 ,通过空间相关性分析 ,寻求从热红外遥感图像中直接分离出大气的空间分布模式的方法 ,以解决地温反演中面临的困难。已有的研究结果说明了在全球变化和定量遥感研究中 ,对地表多尺度空间变化信息作简洁的地学描述的必要性及其重要的应用意义。

The global change study requires information about major factors for spatio\|temporal heterogeneity of land surface, which can be obtained by quantitative remote sensing. This paper presents some new idea and mathematic approach on effective geographical descriptions for land surface feature. Based on the analysis for the limitation of variogram, the conception of histo\|variogram is presented as a way to find the spatial distribution characteristics for typical land surface, by defined standing pixel and contour pixel in the processing of scaling\|up. This is demonstrated successfully by an experimental example. Further more, such spatial distribution information gained from visible and near\|infrared remote sensing images can be used in thermal remote sensing, to retrieve land surface temperature. The thermal remote sensing image can be separated into two parts, the contribution of land surface and the contribution of atmosphere (SAS), based on the different spatial distribution patterns between them. By correlation analysis for different objections of land surface, the method to separate the atmosphere spatial distribution pattern directly from thermal remote sensing image was developed, which should be a possible way to solve the difficult problem in thermal remote sensing. Our new approaches show the necessity and significance of clear and terse geographical description on multi\|scales information about land surface feature in quantitative remote sensing.