基于Sentinel光学和雷达遥感影像的泥炭沼泽识别方法研究

泥炭沼泽是主要的湿地类型之一，准确、快速地获取泥炭沼泽空间分布信息是泥炭沼泽保护、利用和管理的重要基础。然而，泥炭沼泽以土壤中的泥炭层为标志，其表面覆盖的植被类型复杂多样，基于遥感技术精准提取泥炭沼泽的空间分布信息是湿地遥感研究的国际热点问题。本项目以密集时间序列的多光谱（Sentinel-2）和雷达（Sentinel-1）卫星影像为数据源，借助野外调研、实地光谱测量、雷达探测和地形等数据，揭示泥炭沼泽的光谱和雷达后向散射特征，探究泥炭沼泽识别和提取的遥感机理；利用GEE云平台和随机森林算法，实现大兴安岭地区泥炭沼泽的识别与提取，并对分类精度进行验证。本项目的研究成果可为其它地区泥炭沼泽遥感制图研究提供方法借鉴，并可为泥炭沼泽的管理、保护与可持续利用提供数据基础和科学依据。

Peatland is one of the most important wetland types. It is essential to identify the distributions of peatlands accurately and rapidly for their development, utilization, and management. Peat layer is the basic sign determining peatland, however, it is challenging to accurately extract the spatial distribution of peat swamp based on remote sensing technology due to the complex and diverse types of vegetation cover on the surface. In this project, high resolution and dense time series radar Sentinel-1 and optical Sentinel-2 remote sensing data will be used as data sources. The spectral and radar backscattering characteristics of peatlands will be revealed by the data of field spectral surveys, radar detection, and topographic data, to explore the remote sensing mechanism of peatland recognition and extraction. The random forest algorithm will be employed to identify and depict peatland in Daxing'anling area, and the classification accuracy will be verified by field investigations. The research results could be used as reference for the study of peatland mapping in other areas and could provide data basis and scientific basis for the management, protection, and ecological evaluation of peatland.