基于无人机高光谱的盐城滨海湿地恢复区互花米草制图  

UAV hyperspectral mapping of Spartina alterniflora in Yancheng coastal wetland restoration area

作　　者：唐希颖 窦志国 赵欣胜[1] 王俊杰 翟夏杰 李伟[1] TANG Xiying;DOU Zhiguo;ZHAO Xinsheng;WANG Junjie;ZHAI Xiajie;LI Wei(Institute of Wetland Research,Chinese Academy of Forestry,Beijing Key Laboratory of Wetland Ecological Function and Restoration,Beijing 100091,China;Shenzheng University,College of Life sciences and Oceanography,MNR Key Laboratory for Geo–Environmental Monitoring of Great Bay Area,Shenzhen 5180604,China)

机构地区：[1]中国林业科学研究院湿地研究所,湿地生态功能与恢复北京市重点实验室,北京100091 [2]深圳大学生命与海洋科学学院,自然资源部大湾区地理环境监测重点实验室,深圳518060

出　　处：《生态科学》2025年第1期108-118,共11页Ecological Science

基　　金：黄海湿地研究院课题项目基金资助项目(20210109);国家自然科学基金青年项目(42101308);国家重点研发计划项目(2017YFC0506200)。

摘　　要：无人机高光谱遥感是湿地植物物种识别和分类的重要手段,且基于湿地植物光谱特征差异的采用高光谱遥感影像实现对入侵物种的监测具有重要意义。研究以江苏盐城滨海湿地恢复区为研究区,采用无人机高光谱遥感影像作为数据源,结合野外调查获取采样点位的原始光谱曲线,并作一阶微分和连续统去除处理,通过求算均值置信区间和单因素方差分析选择特征波段和特征植被指数。组合光谱特征和植被指数特征构建特征集,基于支持向量机(SVM)、最大似然比(MLC)、马氏距离(MD)方法实现对研究区内互花米草(Spartina alterniflora)生长面积的提取。结果表明,利用变换后的光谱特征和植被指数特征的差异能够较好地实现江苏盐城滨海湿地恢复区内入侵物种的实时监测;整体上,互花米草分布从海水区向半围封区表征为由广密到稀疏,其中,采用SVM、MLC更能较好地实现对研究区内互花米草生长面积的提取,识别精度均为0.89。研究可为利用无人机高光谱实现滨海湿地入侵物种分布监测和制图提供一定的理论依据。UAV hyperspectral remote sensing is an important means to identify and classify wetland plant,and the use of hyperspectral remote sensing images based on the spectral characteristics of different wetland plants is of great significance for the monitoring of invasive species.In this study,the coastal wetland restoration area of Yancheng,Jiangsu Province was taken as the research area,and the hyperspectral remote sensing image of unmanned aerial vehicle was used as the data source.Feature bands and feature vegetation indices were selected by calculating mean confidence intervals and one–way ANOVA.A feature set was constructed by combining spectral features and vegetation index features,and based on Support Vector Machine(SVM),Maximum Likelihood Classifier(MLC),and Mahalanobis Distance(MD)methods to achieve the growth area of Spartina alterniflora in the study area.The results showed that the real–time monitoring of invasive species in the coastal wetland restoration area of Yancheng,Jiangsu could be better realized by using the difference of the transformed spectral characteristics and vegetation index characteristics.The extraction of the growth area of Spartina alterniflora in the study area was achieved well,and the recognition accuracy was 0.89.The research can provide a certain theoretical basis for monitoring and mapping the distribution of invasive species in coastal wetlands using UAV hyperspectral.

关 键 词：无人机高光谱 光谱特征 入侵物种 物种识别 支持向量机 最大似然分类 马氏距离 

分 类 号：X513[环境科学与工程—环境工程]