提出一种基于数学形态学的空间信息和光谱信息集成的影像分类方法,并分析该方法相对于单纯基于像元光谱特征的分类方法所具有的优势。对ROSIS高分辨率高光谱影像、北京中关村地区OMIS高光谱遥感影像两组数据进行SVM分类验证,试验表明,基于空间信息和光谱信息的分类技术能产生较好的分类结果。 An image classification method integrating spatial information and spectral information based on mathematical morphology is proposed and the advantages of this method over the classification method based on pixel spectral features are analyzed. Two groups of data of ROSIS high resolution hyperspectral image and OMIS hyperspectral remote sensing image of Zhongguancun area in Beijing were verified by SVM. The experiments show that the classification technology based on spatial information and spectral information can produce better classification results.