正交子空间投影(OSP)方法广泛用于目标与背景的分离之中,对于高光谱影像,OSP可用于目标提取和混合像元分解,但缺点是需要端元的先验知识。针对这一问题,本文基于OSP的原理提出了一种非监督快速端元提取方法。实验使用模拟高光谱数据和由OM ISⅠ获取的真实高光谱数据,结果精度令人满意,证明了本文算法进行端元自动提取的可行性。 Orthogonal Subspace Projection (OSP) method is widely used in the separation of target and background. For hyperspectral images, OSP can be used for target extraction and mixed pixel decomposition, but the disadvantage is that it requires prior knowledge of endmember. In response to this problem, this paper proposes a non-supervised fast endpoint extraction method based on OSP principle. Experiments using the simulated hyperspectral data and the real hyperspectral data acquired by OM ISⅠhas satisfactory results with satisfactory accuracy. It proves the feasibility of this algorithm for automatic extraction of endmember.