基于植被物候特征的互花米草提取方法研究——以长三角湿地为例

互花米草自1979年引进我国以来迅速增长,呈现外来物种入侵势态,严重影响了滨海湿地生态系统平衡。通过遥感手段监测可以获取互花米草的时空扩展规律,为互花米草治理提供参考和依据。本文选取长三角地区3个主要湿地区域为研究区,依托Google Earth Engine (GEE)平台选取了2014—2019年Landsat 8 OLI时间序列数据,提出了一种基于物候特征的互花米草提取方法。首先,通过归一化差异湿度指数(NDMI)、归一化差异植被指数(NDVI)和归一化差异水体指数(NDWI)提取互花米草生长的高湿度区域;然后,通过植被指数构建表征植被物候特征的时间序列曲线,确定互花米草与其他植被的物候特征差异时相;最后,基于物候特征差异时相数据构建决策树提取互花米草。通过与现有的互花米草决策树提取方法和支持向量机(SVM)方法对比发现,本文方法在3个研究区的提取结果均为最优提取结果,表明本文的方法对提取互花米草具有较好的适用性。

Extraction method of Spartina alterniflora based on vegetation phenology characteristics: a case study of wetlands in the Changjiang River Delta

Spartina alterniflora has been rapidly growing in China since it was introduced in 1979, presenting a trend of invasion of alien species, which seriously affects the ecosystem balance of coastal wetlands. The spatial and temporal expansion of Spartina alterniflora can be obtained by remote sensing monitoring, which provides reference and basis for Spartina alterniflora management. In this paper, three major wetland areas in the Changjiang River Delta were selected as the research area.Based on the Google Earth Engine (GEE) platform, Landsat 8 OLI time series data from 2014 to 2019 were selected, and an extraction method of Spartina alterniflora based on phenological characteristics was proposed. First, the normalized differential moisture index (NDMI), normalized differential vegetation index (NDVI) and normalized differential water index (NDWI)were used to extract the high moisture areas of Spartina alterniflora. Then, the time series curves representing the phenological characteristics of vegetation were constructed by vegetation index, and the time of the data with the difference in phenological characteristics between Spartina alterniflora and other vegetation was determined. Finally, a decision tree was constructed to extract Spartina alterniflora based on temporal data of phenological characteristics differences. Compared with the existing decision tree extraction method of Spartina alterniflora and support vector machine (SVM) method, the extraction results of the proposed method in the three research areas are all optimal, which indicates that the proposed method has a good applicability for the extraction of Spartina alterniflora.