首页> 外文会议>Lidar technologies, techniques, and measurements for atmospheric remote sensing XIII >A Graph Signal Filtering Based Approach for Detection of Different Edge Types on Airborne LiDAR Data
【24h】

A Graph Signal Filtering Based Approach for Detection of Different Edge Types on Airborne LiDAR Data

机译:基于图信号滤波的机载LiDAR数据不同边缘类型检测方法

获取原文
获取原文并翻译 | 示例

摘要

Airborne Laser Scanning is a well-known remote sensing technology, which provides a dense and highly accurate, yet unorganized point cloud of earth surface. During the last decade, extracting information from the data generated by airborne LiDAR systems has been addressed by many studies in geo-spatial analysis and urban monitoring applications. However, the processing of LiDAR point clouds is challenging due to their irregular structure and 3D geometry. In this study, we propose a novel framework for the detection of the boundaries of an object or scene captured by LiDAR. Our approach is motivated by edge detection techniques in vision research and it is established on graph signal filtering which is an exciting and promising field of signal processing for irregular data types. Due to the convenient applicability of graph signal processing tools on unstructured point clouds, we achieve the detection of the edge points directly on 3D data by using a graph representation that is constructed exclusively to answer the requirements of the application. Moreover, considering the elevation data as the (graph) signal, we leverage aerial characteristic of the airborne LiDAR data. The proposed method can be employed both for discovering the jump edges on a segmentation problem and for exploring the crease edges on a LiDAR object on a reconstruction/modeling problem, by only adjusting the filter characteristics.
机译:机载激光扫描是一种众所周知的遥感技术,它提供了密集且高度准确的但无组织的地球表面点云。在过去十年中,许多地理空间分析和城市监控应用研究都致力于从机载LiDAR系统生成的数据中提取信息。但是,由于LiDAR点云的不规则结构和3D几何形状,因此处理起来具有挑战性。在这项研究中,我们提出了一种新颖的框架来检测LiDAR捕获的物体或场景的边界。我们的方法受到视觉研究中边缘检测技术的启发,并建立在图形信号滤波的基础上,该图形信号滤波是对不规则数据类型进行信号处理的令人兴奋且有希望的领域。由于图形信号处理工具在非结构化点云上的便捷适用性,我们通过使用专门为满足应用程序要求而构造的图形表示来实现在3D数据上直接检测边缘点。此外,将海拔数据视为(图形)信号,我们利用了机载LiDAR数据的空中特性。所提出的方法既可以用于发现分割问题上的跳变边缘,也可以通过仅调整滤波器特性来研究重建/建模问题上的LiDAR对象上的折痕边缘。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号