结合高分辨率多光谱影像和Lidar数据提取城区建筑物.pdfVIP

结合高分辨率多光谱影像和LiDAR数据提取城区建筑物 1 2 谭衢霖 ，王今飞 1 北京交通大学土木建筑工程学院，北京 100044 ；2 Department of Geography, The University of Western Ontario, London, N6G 5C2, Canada 摘 要：本文结合 IKONOS 影像和 LiDAR 数据，应用面向对象分类分析方法试验提取高分 辨率多光谱卫星影像城区建筑物。处理流程包括如下步骤：（1）影像融合增强；（2 ）影像分 割；（3 ）影像对象分类；（4 ）建筑物对象几何形状规则化处理。试验结果表明，面向对象分 类分析是一种适于利用高空间分辨率遥感影像数据进行大比例尺城区细部结构制图的有效 方法。该方法对于快速获取城区建筑物分布和辅助制订城区发展规划具有应用价值。 关键词：高分辨率；面向对象；分割；分类 中图分类号：P407. 8 文献标识码：A 文章编号： Building Extraction Using High Resolution Multi-spectral Image and Lidar Data 1 2 Tan Qulin , Wang Jinfei 1.School of Civil Engineering, Beijing Jiaotong University, Beijing, 100044; 2. Department of Geography, The University of Western Ontario, London, N6G 5C2, Canada Abstract: In the paper, the object-oriented image analysis method was applied to building roof mapping using IKONOS multi-spectral images combined with Lidar data. The scheme to produce the vector building-polygons includes the following steps: 1 image fusion and derivation of ancillary input data, 2 segmentation of processed data layers into image objects, 3 classification of image objects, and 4 geometrical regularization of classified building object polygons. The experimental results show that object-oriented classification analysis is an effective method for large-scale detailed urban structure mapping using very high-resolution remote sensing images. The procedure may suit overall investigation of urban building development and planning. Keywords: High resolution; Object-oriented; Segmentation; Classification 收稿日期： 修订日期： 基金项目：国家自然科学基金项目（座机电话号码, 座机电话号码 ）；中央高校基本科研业务费专项项目（2009JBM067 ） 作者简介：