基于无人机高光谱遥感的典型草原退化进程监测及评价研究

Unmanned aerial vehicle (UAV) hyperspectral remote sensing can obtain low-altitude and high-resolution hyperspectral remote sensing data. It can realize timely, fast and time-saving high-resolution dynamic monitoring for grassland degradation, making up a great amount of manpower and material resources consumption in ground survey, regional real-time monitoring difficulties, lower accuracy of satellite remote sensing, difficulties in dynamic quantitative analysis, which will become an important means for dynamic monitoring and quantitative research of grassland degradation. In this study, taking grassland vegetation and soil as the research object, with cover, biomass, community structure, bare land, rat holes, sand patch etc. as evaluation index, dynamic monitoring for month period, collecting low-altitude UAV hyperspectral remote sensing data, using ground sampling for precision inspection in accuracy of identification and classification, obtaining the corresponding index of grassland degradation in highspectral remote sensing background data, establishing the hyperspectral remote sensing model to grassland degradation of typical steppe. Hyperspectral remote sensing classification and degree of degradation in plant community structure of typical steppe will be tested by constructing the relationship between identification of surface micro patch and grassland degradation in typical steppe, and ultimately we will provide precise, dynamic, quantitative and real-time technology for regional monitoring of grassland degreadation, at the same time this study will enrich scientific theory of grassland degradation, and also provides basis for effect evaluation of grassland ecosystem recovery and reconsturction.

无人机高光谱遥感能获取低空高分辨率的高光谱遥感数据，实现对草原退化及时、快速、省时的高分辨率动态监测，弥补了地面调查人力物力消耗大,区域性实时监测困难、卫星遥感精度较低、难以进行动态定量分析的弊端，因而将成为草原退化动态监测与定量研究的重要手段之一。本研究以实验区草原植被和土壤为研究对象，选择盖度、生物量、群落结构、裸地、鼠洞、沙斑等指标，以月为周期进行动态监测，采集无人机低空高光谱遥感数据，用地面样方对识别精度和分类精度进行精度检验，获得典型草原退化相应指标的高光谱本底数据，建立典型草原退化的高光谱遥感数据模型，研究典型草原植物群落结构的高光谱遥感分类与退化程度，建立典型草原地表微斑块的识别及与草原退化的关系，为草原退化提供高精度、动态、定量、实时的技术，丰富草地退化的科学理论，为草原系统恢复与重建效果的评价提供依据。

草原退化是中国主要的生态环境问题之一。本研究开展了基于无人机高光谱遥感的草原退化研究，按照研究计划，分别完成了以下工作:构建了两套不同类型的无人机高光谱遥感系统，为实现高空间分辨率、多光谱维度信息，且实时、高效的草原退化监测与研究提供了技术手段；连续三年以月份为周期进行草原实地数据采集，获得了退化草原的高光谱研究基础数据；研究了草原原始种、优势种和退化指示种等草原草种类在开花期、结实期、枯黄期的光谱特性，并建立了《植物类型高光谱特征数据库》，揭示了草原植物光谱特性随生长期的变化规律；研究了鼠洞等草原地表微斑块的光谱特性，并建立了《地表微斑块高光谱数据库》；利用深度学习数据分析手段并结合多数据源融合技术，分别建立了草原植被特征和地表特征的分类、识别模型，实现了基于无人机高光谱遥感数据的草原地物分类与识别，为基于无人机高光谱遥感的大面积草原退化研究提供了监测数据与数据分析方法。上述工作，为下一步进行基于无人机高光谱遥感的草原退化评价体系研究奠定了基础。项目资助发表SCI论文2篇，待发表1篇，EI论文1篇，中文核心论文5篇，待发表1篇，获得实用新型专利1项，软件著作权2项，培养博士研究生5名，其中1名已经取得博士学位，4名在读；培养硕士研究生8名，其中6名已毕业，2名在读。项目投入经费40万，支出23.2586万元，各项支出基本与预算相符。剩余经费16.7414万元，计划用于本项目后续研究支出。