无瓣海桑作为人工引进的外来红树物种是否具有生态入侵性存在较大的争议。该物种已经广泛地分布于我国主要的红树林区域，但由于中高分辨率遥感数据难以体现植被间的差异，无瓣海桑在我国准确的空间分布信息还未有报道。在复杂的滨海环境下，如何利用遥感技术精确地获取大范围无瓣海桑的空间分布信息是亟待解决的科学问题。本项目以中国红树林作为研究对象，集成时序Sentinel雷达和光学影像，提出一种结合植被物候特征和深度学习的物种信息提取方法。该方法能够充分挖掘时序遥感影像在时相域、空间域和波谱域的深层特征信息，解决中高分辨率影像在物种分类中高判别性的关键问题，并将致力于首次获取中国沿海无瓣海桑的空间分布信息。本项目的实施对于大范围植被物种水平的精细化遥感监测具有重要的科学意义，对于提升红树林植被监测能力、评估人工修复物种的生态效应和选择造林树种等都具有重要的应用价值，能够为无瓣海桑引种存在的争端提供数据支持。

Alien mangrove species, Sonneratia apetala (SA), was introduced to artificial restoration. The issue of the biological invasion of this species has raised great debates recently. The species has been widely distributed in all main mangrove areas in China. However, the spatial distribution pattern of SA in China is still unknown as the medium-high images are challenging to distinguish mangrove species. Under the complex coastal environment, how to accurately monitor alien mangrove temporal and spatial changes over time need further research. The mangrove forests in China are selected as the study area. We develop the model by integrating plant phenological trajectory features and deep learning to achieve species identification across a wide range. The model can be used to mine spatial-temporal-spectral features of multi-source time-series images to improve species classification accuracy. The model can be applied to identify SA accurately. The results are meaningful to monitor a wide range of vegetation species. They will improve the monitoring capacity of mangrove vegetation and serve as a baseline to assess the effect of artificial restoration mangrove. The resultant datasets will be used to resolve the debates of SA.