多源多尺度DEM互增强方法研究

To make full use of the complementary advantages among DEMs with various resources and resolutions, the co-enhancement of DEMs with multi-resources and multi-resolutions is aimed to improve both the resolution and accuracy of DEM. Facing the requirements of high accuracy and high resolution for regional and global scales study, this research intends to systematically construct a new theoretical approach and framework structure of co-enhancement of DEMs with multi-resources and multi-resolutions, taking the existing global-scale DEMs as its data resources. Based on lots of theoretical study and experimental analysis, the core research contents of this project contain the following four aspects. Firstly it aims to study a new method for the priori quality assessment of DEMs with multi-resources and multi-resolutions, according to factors such as data acquisition technology and DEM construction methods etc. Secondly, the study mission is to construct a resolution calculation model for DEMs with multi-resources and multi-resolutions, concerning the complementarity among these various DEMs. Moreover, on account of the constraints of topography structure, a spatial interpolation model adapted to multi- variables and variable kernel functions is to be built here. Based on the whole enhancement process and the above work, the last task of this project is to set up a quality assessment model for enhanced DEMs and develop a processor for co-enhancement of DEMs from various data sets. Through this research, the fusion of DEM elevation will be promoted to a higher level of co-enhancement both on the DEM resolution and elevation. What is more, the fusion between two DEMs will also be advanced to the level of co- enhancement among multiple DEMs. In conclusion, our work can not only facilitate the comprehensive utilization of DEMs with global scales, but also has great theoretical and practical meanings on expanding the wide area of DEM applications and improving the theories and methods of DEM.

多源多尺度DEM互增强是通过不同分辨率和不同精度DEM之间的优势互补，实现DEM在分辨率和精度上的提高。面向区域、全球尺度研究和应用等对高精度、高分辨率DEM的需求，本研究以已有全球尺度DEM为数据源，在理论研究和实验分析的基础上，系统研究多源多尺度DEM互增强的理论方法和实现框架。从已有DEM的数据获取、生成机理等方面，探索多源多尺度DEM的先验质量评估方法；在考虑多源多尺度DEM的互补信息基础上，构建多源多尺度DEM互增强的分辨率计算模型；考虑地形形态结构约束，研究多变量可变核自适应空间插值模型；从增强处理过程入手，构建增强DEM的质量评价模型；研发DEM互增强处理器。研究将DEM从高程融合层面推向分辨率和高程同时增强的层面，从两个DEM融合推向多个DEM互增强的层面。研究对全球尺度DEM的综合利用、拓展DEM应用广度、完善DEM理论与方法具有重要的理论和实践意义。

目前各类开源 DEM 由于数据获取与处理方式不同，在精度和空间分辨率上存在较大差异，实现不同 DEM 在分辨率上的提升和精度上的增强是有望改善单一DEM的质量。本研究基于开源DEM，提出了DEM增强的数学模型和概念模型，构建了DEM增强的框架体系结构；从地形参数、地形结构和地表覆盖层面，较为系统地分析了开源DEM的误差特征和空间分布，形成了开源DEM可靠区域的判断策略；研究了基于分形维的增强DEM分辨率计算方法，考虑到目前DEM增强定权存在问题，通过引入可变核的非参数内插方法构建了多源多尺度DEM互增强模型，有效解决多个DEM相互融合的问题；研究了DEM形态的质量评估方法，设计了DEM地形形态描述指标，从数据和形态两个角度构建了增强DEM的评价体系。研究对全球尺度 DEM 的综合利用、拓展 DEM 应用广度、完善 DEM 理论与方法具有重要的理论和实践意义