草地系统认知理论与光谱融合量化策略

研究目的：面对全球气候变化、社会经济发展以及粮食结构转型，运用先进的遥感技术探索草地系统规范的认知理论和遥感解译框架。研究方法：通过文献分析发现标准光谱端元空间或是突破传统遥感技术壁垒实现草地系统认知规范的关键；基于光谱混合分解模型构建多光谱—高光谱两个尺度的标准光谱端元空间，并开展草原类型多级嵌套分类实验。研究结果：提出草地系统认知理论框架和基于标准光谱端元空间的光谱融合量化策略，并在中国—内蒙古—科尔沁左翼后旗三个尺度的草地类型多级嵌套遥感分类实验中取得较好的结果。研究结论：基于标准光谱端元空间的光谱融合量化策略有利于构建稳定、具有专业概念支撑的解译框架，配合多级嵌套的分类结构能够实现草地系统认知框架下光谱数据的规范，从而支撑不同等级的应用和管理需求。

Grassland System Cognitive Theory and Its Spectral Identification Method

The purpose of this paper is to absorb the latest grassland resource theory and modern remote sensing technology to establish a monitoring and evaluation framework of grassland utilization and management system, under the dual pressure of increasing service demand and ecological degradation. The research methods are as follows. This paper constructs the grassland system cognitive theory based on the four-dimensional scientific theory from the three aspects of morphological structure, energy balance and behavioral information. At present, the rapid development of remote sensing technology provides an opportunity to update the investigation methods of grassland resources and the concretization of grassland system cognitive theory. At the same time, the transformation from surface mixed spectral space to surface terminal space can break through the dilemma faced by the development of remote sensing technology. Based on the theory of grassland cognitive system, this paper constructs a quantitative strategy of grassland system spectral fusion based on standard spectral endmember space. The research results show that multi-level nested remote sensing classification of grassland types is applied from three scales: potential vegetation zoning of grassland in China with products of MOD11A2, MOD13A3, TRMM 3B43 and DEM; classification of grassland types in Inner Mongolia with MOD09A1 and classification of grassland classes, groups and types in Horqin Left Back banner with hyperspectral data of ziyuan-1 02D satellite. In conclusion, the spectral fusion quantitative strategy of grassland cognitive system based on endmember space can realize the dynamic monitoring of grassland system evolution process and mechanism at different scales. It demonstrates the potential of this study for the transformation of grassland resources investigation, and provide important support for grassland resource management.