Classification and Change Detection in Mobile Mapping LiDAR Point Clouds

Abstract

Creating 3D models of the static environment is an important task for the advancement of driver assistance systems and autonomous driving. In this work, a static reference map is created from a Mobile Mapping “light detection and ranging” (LiDAR) dataset. The data was obtained in 14 measurement runs from March to October 2017 in Hannover and consists in total of about 15 billion points. The point cloud data are first segmented by region growing and then processed by a random forest classification, which divides the segments into the five static classes (“facade”, “pole”, “fence”, “traffic sign”, and “vegetation”) and three dynamic classes (“vehicle”, “bicycle”, “person”) with an overall accuracy of 94%. All static objects are entered into a voxel grid, to compare different measurement epochs directly. In the next step, the classified voxels are combined with the result of a visibility analysis. Therefore, we use a ray tracing algorithm to detect traversed voxels and differentiate between empty space and occlusion. Each voxel is classified as suitable for the static reference map or not by its object class and its occupation state during different epochs. Thereby, we avoid to eliminate static voxels which were occluded in some of the measurement runs (e.g. parts of a building occluded by a tree). However, segments that are only temporarily present and connected to static objects, such as scaffolds or awnings on buildings, are not included in the reference map. Overall, the combination of the classification with the subsequent entry of the classes into a voxel grid provides good and useful results that can be updated by including new measurement data.

A Review of Mobile Mapping Systems: From Sensors to Applications

“…An example of these systems is introduced in [73] Vehicle-mounted systems are used for various applications such as urban 3D modeling, road asset management, and condition assessment [78,79]. Moreover, these systems can be used for automated change detection in the mapped regions [80,81], creating upto-date HD maps as an asset for autonomous driving [73] and railway monitoring applications [82].…”

Efficient 3D Mapping and Modelling of Indoor Scenes with the Microsoft HoloLens: A Survey

“…A straightforward solution towards more efficient data acquisition is represented by a Mobile Laser Scanning (MLS) system or a Mobile Mapping System 1 (MMS), since all acquired data are directly co-registered on-the-fly. Such sensor systems are meanwhile commonly used for acquiring the geometry of both outdoor scenes (Paparoditis et al 2012;Gehrung et al 2017;Roynard et al 2018;Voelsen et al 2021) and indoor scenes (Otero et al 2020). Regarding data acquisition within indoor scenes, different solutions are conceivable such as trolley-based systems (e.g., Fig.…”

Change Detection of Urban Trees in MLS Point Clouds Using Occupancy Grids

“…The aim is similar to ours, yet due to the differences in data characteristics, not all previously mentioned methods can be successfully transferred to mobile laser scanning. Voelsen et al (2021) entirely rely on MLS data while detecting changes in the built environment by separating them into static and non-static objects. They identify the effects and interpretation of vegetation changes as problematic, primarily due to the semi-static behaviour of vegetation-the objects are not as rigid over time as, for example, buildings.…”