融合LiDAR点云和高分辨率影像的地物分类研究

Landmark classification by fusing LiDAR and high-resolution remotely sensed image

LiDAR点云包含丰富的高程信息,而高分辨率航空影像则富含光谱信息。在分析两类数据特征优势的基础上,以模糊集为融合方法的理论基础,研究了一种结合LiDAR点云和高分辨率影像的分类方法。该方法首先通过航空影像获取模糊分类结果,然后引入LiDAR点云的高程信息,再进行模糊决策融合,重点改善了建筑物与其他地物之间的混分现象。实验结果表明,在LiDAR点云的辅助下,该方法能够有效改善光谱信息分类中出现的建筑物与裸地、道路、水体等地类之间的混分现象,明显提高了总体分类精度,为复杂城市区域提供了更为精确的地物分类结果。

LiDAR point cloud contains abundant elevation information,and high-resolution aerial image contains rich spectral data.On the basis of the analysis of the advantages of the two types of data,the combination of LiDAR point cloud and high-resolution remotely sensed image was presented using fuzzy set as the theoretical basis of incorporation method.The fuzzy classified results are obtained by aerial image first,which was then in fuzzy decision incorporation by introducing the elevation information of LiDAR point cloud.The results show that the presented method can improve the confused classification among buildings,bare lands,roads and water bodies,which significantly improved the general classification precision and provided more precise landmark classification results for complicated urban area.