一种基于面向对象辅以DSM的建筑物提取方法

摘要

Based on the object-oriented technology and DSN method, the building information is extracted quickly. The slope information is extracted from DSM generated by Li DAR point cloud data, and the outline of building is extracted according to the height information of object. The object outside the outline of the building is then subdivided. The outline of the building is optimized based on the spatial information, shape information and spectral information of the divided building objects. The experimental results show that the correct rate of the extracted objects is 97. 9％, and the completeness rate is 96. 4％.%基于面向对象技术并辅以DSN的方法来快速提取建筑物信息, 从LiDAR点云数据生成的DSM中提取到坡度信息, 并根据对象的高度信息来提取到建筑物的大致轮廓, 然后分割建筑物轮廓外的对象, 依据已分建筑物对象的空间信息、形状信息及光谱信息来优化建筑物轮廓.实验结果表明, 提取到的对象的正确率为97. 9％, 完整率为96. 4％.