基于高分辨率遥感影像的湿地互花米草-芦苇混合交错带提取方法

互花米草是中国东部河口滩涂湿地主要入侵物种之一,其与本地物种芦苇竞争生长,形成了大范围的混合交错带。该交错带是研究湿地生态系统动态变化的重要信息,但因2种物种光谱的相似性及其在交错带的组成复杂性,使利用遥感技术提取交错带难度较大。因此,本文提出了一种将二者生长物候差异与其光谱特征相结合,考虑二者海陆位置分布差异,运用实测剖面观测数据确定光谱指标和阈值的综合提取方法。运用高分一号多光谱遥感数据,通过分析不同时相互花米草与芦苇冠层光谱差异,确定用来提取混合交错带的高分遥感影像,实现了研究区互花米草-芦苇混合交错带的提取。结果表明：不同时相宜选用不同的提取指标,本研究中在春季选择了近红外波段反射率,而秋季则选择了红波段反射率;2个时相的混合交错带范围存在明显差异,客观反映了互花米草与芦苇在不同季节的竞争状况。

Extracting the Transitional Zone of Spartina alterniflora and Phragmites australis in the Wetland Using High-resolution Remotely Sensed Images

Spartina alterniflora is one of the major invasive species putting a high pressure to the native Phragmites australis in the coastal wetland in Chongming Island, Shanghai. Both species grow up together and result in extensive transitional zones. Spartina alterniflora and Phragmites australis have differences in physiology. The former prefers to live in a high-salt environment and it distributes closer to the sea. In the transitional zones, the two species mix in different proportion along the direction from the sea to the land and grow competitively with each other. Their growth in the transitional zones reflects the intensity of competition. Moreover, the change of the location of transitional zones reflects the dynamic process of the invasion of Spartina alterniflora. Thus, the transitional zone plays a key role in the study of dynamic change of wetland ecosystem. However, it is difficult to extract such information precisely by remote sensing because of the similar spectra of two species and complex composition of the transitional zone. They have similarities in both spectra and physiological and ecological characteristics because both of them are gramineous plants. In addition, the two species in transitional zones mix with different proportions in different regions, so the composition of the transitional zones is complex. There is little related research focusing on the transitional zones so far. This paper presents a comprehensive extraction method. Firstly, we combine different phenology with spectral characteristics to narrow the scope of the appropriate indicators down to reduce the workload. Secondly, we also consider location difference of the land and the sea. Analyzing the spectral characteristics along the direction from the sea to the land, it will be more intuitive for the spectral characteristics of vegetation in the transitional zones as well as the relationship between the change of mixing ratios and spectral characteristics. Finally, we determine extracting indicators and threshold by actual measured dataset. Remote sensing is an important measure for monitoring wetland ecosystem. Extracting transitional zone requires high-resolution remotely sensed images because the width of transitional zone in our research is narrow and transitional zone contains many patches which is due to many complex factors such as underwater micro-topography differences. This study selected appropriate multi-spectral remotely sensed data from GF-1 as research object through analyzing the canopy spectral differences between Spartina alterniflora and Phragmites australis in different time. Also, this study extracted transitional zone successfully in the study area which is an intertidal area located in the northeastern part of Chongming Island. The results indicate that different indicators should be used in different time. We select near-infrared band reflectance for spring and red band reflectance for autumn The near-infrared band reflectance of vegetation in transitional zone is lower than the other two pure species regions in spring while the red band reflectance is higher than the other two pure species regions in autumn. The scale of transitional zone has obvious difference in the two growing stages, the width of transitional zone in autumn is narrower than that in spring and the location moves towards the direction of Phragmites australis regions. The difference reflects the competitive situation in different seasons, objectively. Competitive edge of Spartina alterniflora is more evident in autumn.