介绍了高光谱图像的数据特点,描述了非监督分类和监督分类算法,以及评价分类效果的方法。利用森林场景高光谱图像,对K值聚类算法、最小距离分类算法和光谱角度填图算法进行了实验验证,并对不同算法进行了评价。试验结果表明,监督分类,尤其是光谱角度填图方法的效果要明显好于非监督分类方法。 The characteristics of hyperspectral image data are introduced. Unsupervised classification and supervised classification algorithms are described, and the method of evaluating the classification results is described. The K-value clustering algorithm, the minimum distance classification algorithm and the spectral angle mapping algorithm are experimentally verified by using hyperspectral images of forest scenes. Different algorithms are evaluated. The experimental results show that the effect of supervised classification, especially spectral angle mapping, is obviously better than unsupervised classification.