论文部分内容阅读
针对已有的从机载激光雷达(LiDAR)点云提取建筑物的方法多需要设置阈值及分类规则,造成算法适应性不强的问题,该文提出了一种LiDAR点云和多光谱影像进行自动化建筑物检测的方法。首先通过数据预处理从LiDAR点云中分离出建筑物点和树木点,然后综合LiDAR点云的表面曲率、强度信息和对应多光谱影像的NDVI值构建特征向量,最后基于支持向量机完成自动化的建筑物检测。试验结果表明,基于支持向量机的方法可将两种数据源有效结合起来用于自动化的建筑物检测。
In order to solve the problem that the existing methods to extract buildings from point cloud of Airborne Lidar (LiDAR) need to set thresholds and classification rules, which results in poor adaptability of the algorithm, a LiDAR point cloud and multi-spectral image are proposed Automated building inspection methods. Firstly, the building points and tree points are separated from the LiDAR point cloud by data preprocessing. Then, the surface curvature and intensity information of LiDAR point cloud and NDVI values of the corresponding multispectral image are used to construct the eigenvector. Finally, the support vector machine Building inspection. The experimental results show that the SVM-based approach can effectively combine the two data sources for automated building inspection.