利用高空间分辨率影像中地物的几何特征和空间邻域信息,结合我国土地利用状况和空间分布的特点,在支持向量机分类方法的基础上,利用几何和图论特征实现高分辨率遥感影像土地利用状况分类,并分析讨论了几何和图论特征对分类结果的影响。实验结果表明,该方法可以有效地识别未利用地、农业用地、工业用地和商住用地4种典型地类,其分类结果也可为土地利用变化监测提供依据。 Based on the features of land use and spatial distribution in China, based on the classification of support vector machines (SVMs) and using the geometry and graph theory features of high spatial resolution imagery to achieve high resolution remote sensing Image classification of land use status, and analyzed the impact of geometry and graph theory features on classification results. The experimental results show that this method can effectively identify four typical types of unused land, agricultural land, industrial land and commercial land, and the classification results can also provide the basis for land use change monitoring.