中文摘要：

碱蓬和柽柳是黄河口湿地典型的盐生植物类型,是多种保护珍禽的主要栖息地,具有景观尺度较小、分布广且多混生的特点。应用覆盖黄河口北部潮滩的HJ-1A高光谱遥感影像,基于现场测量的端元光谱和从遥感影像中使用顺序最大角凸锥法（SAMCC）自动提取的端元光谱,应用线性光谱分解法（LSU）、正交子空间投影法（OSP）、匹配滤波法（MF）、最小能量约束法（CEM）和自适应一致估计法（ACE）5种不同光谱解混方法进行混合像元光谱解混,对比两种方法得到的端元光谱分别对碱蓬和柽柳盖度的反演能力,并给出相应的反演模型。结果显示：（1）现场测量端元光谱取得了较好的碱蓬和柽柳盖度反演结果,其中应用LSU方法的光谱解混结果与现场测量盖度的决定系数对于碱蓬和柽柳分别达到了0.88和0.95;（2）两种端元获取方式的光谱解混结果中,LSU和OSP方法均获得了较高的相关性,ACE解混方法的相关性都最低;（3）SAMCC方法提取端元光谱对柽柳的分解结果与现场测量盖度的相关性远高于碱蓬。

英文摘要：

Suaeda salsa and tamarisk are the typical vegetation in Yellow River Estuary,and also a key habitat for many kinds of rare birds. Suaeda salsa and tamarix,whose landscape scale usually small,widely distributed and of- ten mixed with each other or other vegetation or bare soil in remote sensing images, are difficult to obtain the vege- tation fractional cover efficiently, thereby affecting biomass monitoring and ecological assessment. In this study,we acquired the end-members in two ways, the first is measure the vegetation spectrum through field work, and the second way is extract them from the hyperspectral remote sensing image with the SAMCC method. Then we use the two kinds of end-members in the pixel spectral un mixing process with 5 different algorithms, which are. LUS, OSP,MF,CEM and ACE. At last we evaluate the two kinds of end-members coverage inversion capability of the salsa and Tamarix through correlation analysis with the real coverage which measured in the field work.