A Voxel-Based Metadata Structure for Change Detection in Point Clouds of Large-Scale Urban Areas

Abstract

ABSTRACT:Mobile laser scanning has not only the potential to create detailed representations of urban environments, but also to determine changes up to a very detailed level. An environment representation for change detection in large scale urban environments based on point clouds has drawbacks in terms of memory scalability. Volumes, however, are a promising building block for memory efficient change detection methods. The challenge of working with 3D occupancy grids is that the usual raycasting-based methods applied for their generation lead to artifacts caused by the traversal of unfavorable discretized space. These artifacts have the potential to distort the state of voxels in close proximity to planar structures. In this work we propose a raycasting approach that utilizes knowledge about planar surfaces to completely prevent this kind of artifacts. To demonstrate the capabilities of our approach, a method for the iterative volumetric approximation of point clouds that allows to speed up the raycasting by 36 percent is proposed.

A Representation of MLS Data as a Basis for Terrain Navigability Analysis and Sensor Deployment Planning

“…Based on the theoretical foundation of this work, Gehrung et al (2016) proposed a concept for occupancy representation on a global scale. In addition, an algorithm for iterative refinement has been proposed in order to reduce discretization artifacts inherent to the technology as well as to speed up the generation of the occupancy octree (Gehrung et al, 2018).…”

“…Following a well established practice regarding evidence grids used in works such as the one of Hornung et al (2013), octree cells are considered to be independent of each other. The octree is generated using iterative refinement, such as presented by Gehrung et al (2018), since it is less resource intensive than normal ray casting.…”

“…Therefore, space is sampled along a three-dimensional grid and ray casting is applied between the given location and the center of each grid voxel. Since ray casting is quite slow, the actual implementation is based on a modified version of the iterative refinement algorithm proposed by Gehrung et al (2018). Starting from the root node of the occupancy density octree, only voxels that are intersected by the line-of-sight are further refined.…”

A Fast Voxel-Based Indicator for Change Detection Using Low Resolution Octrees

“…Based on the theoretical foundation of this work, we proposed a concept for occupancy representation on a global scale (Gehrung et al, 2016). Later on, an algorithm for iterative refinement has been proposed that also prevents artifacts caused by discretization inherent to occupancy grids (Gehrung et al, 2018). It was also shown that the probabilistic occupancy grids can be used to extract moving objects (Gehrung et al, 2017).…”

Data Processing and Recording Using a Versatile Multi-Sensor Vehicle

“…With regard to the hardware, we based the design of the vehicle on past experiences with the development of an airborne laser scanning system (Schatz, 2008). Currently we use the vehicle for several research topics in the area of object detection and tracking (Becker et al, 2016;Borgmann et al, 2017Borgmann et al, , 2018Hammer et al, 2018), mobile mapping and change detection (Gehrung et al, 2018). In the later part of this paper, we present the use case of pedestrian detection, which uses several capabilities of the vehicle.…”