【摘要】 城市土地利用分类研究对实现城市土地资源高效管理和城市可持续发展具有重要意义。基于无人机获取的遥感影像数据，提出了一种融合DSM特征、三维纹理特征和改进型植被指数的多模态特征提取技术，并利用支持向量机（SVM）分类器进行城市土地利用分类。针对传统绿蓝植被指数对无人机影像敏感度低的问题，提出了一种改进型绿蓝植被指数（MGBVI）。结果表明，采用多模态特征后分类精度提高了16.5%；多模态特征中选择三维纹理特征和MGBVI进行分类的效果最佳。更多还原

【Abstract】 The research on urban land-use classification is of great significance to realize the efficient management of urban land resources and urban sustainable development. Based on the remote sensing image data obtained by UAV, we proposed a multi-modal feature extraction technology integrating DSM feature, 3D texture feature and modified vegetation index, and used support vector machine(SVM) classifier to classify urban land-use styles. In view of the low sensitivity of traditional green and blue vegetation index to UAV images, we proposed a modified green and blue vegetation index(MGBVI). The results show that the classification accuracy is improved by 16.5% after using multi-modal features, and the 3D texture feature and MGBVI are the best among multi-modal features.更多还原