车载激光扫描数据路坎点云提取方法

发布时间：2018-05-08 05:56

本文选题：车载激光点云 + 支持向量机(SVM)　；参考：《地球信息科学学报》2017年07期

【摘要】：车载激光扫描系统能够快速准确地获取街道环境的点云数据,但由于扫描点云的点密度高、数据量大、空间分布不均匀、地物相互遮挡及城市街道环境复杂等特点,难以直接从原始点云数据中提取出路坎点云。本文首先通过分析路坎点云的空间分布特征和局部几何特征,构建包含相对高程、法向量方向、多尺度高程差及多尺度高程方差的点云特征向量;然后,采用SVM提取城市街道环境车载激光扫描数据中的路坎点云,并对提取结果进行聚类去噪,优化路坎点云。最后,通过Street Mapper 360系统和Lynx Mobile Mapper V100系统采集的4份不同城市街道环境车载激光扫描数据对本文方法进行验证,其中路坎点云提取结果的完整度均超过了94.99%、准确度均超过91.88%、精度亦均达到了90.55%以上。实验结果表明,本文方法能够精确地提取复杂城市街道环境中规则或不规则的路坎点云,且具有较强的稳健性,适用于各类复杂的城市街道环境。
[Abstract]:The vehicle laser scanning system can quickly and accurately obtain the point cloud data of the street environment. However, because of the high point density of the scanning point cloud, the large amount of data, the uneven spatial distribution, the mutual occlusion of the ground objects and the complexity of the urban street environment, etc. It is difficult to extract the outlet point cloud directly from the original point cloud data. In this paper, firstly, by analyzing the spatial distribution and local geometric characteristics of the road ridge point cloud, we construct the point cloud feature vector including relative elevation, normal vector direction, multi-scale elevation difference and multi-scale elevation variance. The road ridge point cloud from the laser scanning data of urban street environment is extracted by SVM, and the results are clustered and denoised to optimize the road ridge point cloud. Finally, the method is verified by four laser scanning data collected by Street Mapper 360 system and Lynx Mobile Mapper V100 system in different urban street environment. The degree of completeness and accuracy of the extraction results are all over 94.99 and 91.88, respectively, and the accuracy is more than 90.55%. The experimental results show that the proposed method can accurately extract regular or irregular road kerb point clouds from complex urban street environments and has strong robustness and is suitable for all kinds of complex urban street environments.
【作者单位】：福州大学地理空间信息技术国家地方联合工程研究中心空间数据挖掘与信息共享教育部重点实验室福建省空间信息工程研究中心;
【基金】：国家自然科学基金青年科学基金项目(41501493) 福建省科技计划重点项目(2015H0015) 福建省教育厅中青年教师科研项目(JAT160078) 中国博士后面上科学基金项目
【分类号】：P225.2;TP391.41

【相似文献】