摘要
森林沼泽作为独特的湿地类型,对区域碳循环和生物多样性保护具有重要作用。以吉林省哈尼湿地为研究对象,集成应用Sentinel-1雷达和Sentinel-2多光谱影像,充分挖掘Sentinel-1多极化波段特征和Sentinel-2红边指数对森林沼泽遥感识别的潜力,应用随机森林方法实现哈尼湿地森林沼泽分布信息的提取。结果表明:当森林沼泽信息提取最优决策树数目为1200,融合Sentinel-1 VV与VH后向散射系数雷达波段特征与Sentinel-2红边波段特征能显著提高森林沼泽分布信息的提取精度,保护区森林沼泽信息提取总体精度与Kappa系数分别高达89%、0.85,提取质量相对较高。哈尼湿地景观类型多样化,天然湿地、人工湿地与非湿地景观类型并存。在天然湿地类型中,森林沼泽面积最大,为27.1 km^(2),占保护区总面积的11.2%;河流面积最小,为0.7 km^(2),占保护区总面积的0.3%。森林沼泽提取方法为哈尼湿地的可持续管理提供数据支持,也为其他地区森林沼泽制图提供案例指导。
Swamp,one of the important wetland types,plays an important role in regional carbon cycle and biodiversity conservation.In this study,we aimed to mapping the swamps in the Hani wetland located in China’s Jilin Province by combining Sentinel-1 radar,Sentinel-2 multispectral images and Random forest algorithm.The potentials of Sentinel-1 multi-polarization band characteristics and Sentinel-2 red edge index were fully investigated for delineating swamps.Results showed that combination of Sentinel-1 VV and VH radar bands and Sentinel-2 red edge band significantly improved the classification accuracy of swamps in the study area when the number of decision trees is 1200.The overall accuracy and Kappa coefficient of swamp is 89%and 0.85,respectively,that suggests a good classification quality.There are obvious heterogeneous landscapes in the protected area including natural and human-made wetlands,as well as non-wetland landscapes.Among the natural wetlands,Swamp had the largest wetland area(27.1 km^(2)),accounting for 11.2%of the total protected area,and the river wetland had the smallest area(0.7 km^(2)),accounting for only 0.3%of the total protected area.The method for delineating swamps provides data support for sustainable management of the Hani wetland,and also provides an example for mapping swamps in other regions.
作者
王景发
毛德华
杜会石
王宗明
WANG Jing-fa;MAO De-hua;DU Hui-shi;WANG Zong-ming(College of Tourism and Geographical Science,Jilin Normal University,Siping 136000,Jilin,China;Key Laboratory of Wetland Ecology and Environment,Northeast Institute of Geography and Agroecology,Chinese Academy of Sciences,Changchun 130102,Jilin,China;National Earth System Science Data Center,Beijing 100101,China)
出处
《湿地科学与管理》
2021年第3期2-7,12,共7页
Wetland Science & Management
基金
国家自然科学基金面上项目(41771383)
中国科学院青年创新促进会人才基金(2017277,2021227,2012178)
吉林省科技发展计划项目(20200301014RQ)。
