An Annotated Multi-Platform Laser Scanning Dataset (MLS, ULS, TLS) for Semantic Segmentation of Point Clouds in Complex Forest Environments
Contributing person | Robert Krüger | |
Contributing person | Hanne Hendrickx | |
Contributing person | Oliver Grothum | |
Contributing person | Isabell Raab | |
Contributing person | Almut Jäschke | |
Additional geographical or spatial references | Saxon Switzerland | |
Countries to which the data refer | GERMANY | |
Description of the data | The dataset provides point clouds of a scene that comprises forested areas with scattered rock formations in the Richterschlüchte area in the southern part of the Saxon Switzerland National Park, Germany; parts of the area were affected by the large-scale wildfires of 2022. It was created within the research project RTG-AirMetro (DFG grant number 508591287) and originates from a field campaign conducted in March 2023. A related manuscript using this data is currently under review. The point clouds were acquired with three laser scanning systems of distinctly different characteristics: a handheld mobile laser scanning (MLS) system operated along a hiking trail leading into the valley, an uncrewed aerial vehicle laser scanning (ULS) system, and a terrestrial laser scanning (TLS) system, from which sections of the nearby ridge are included. The platforms cover the scene from complementary viewing geometries. The ULS and TLS sections overlap only minimally, while the MLS coverage is spatially independent of both, and the MLS subsets are disjoint to prevent data leakage during model development. The dataset consists of five point clouds: the MLS point cloud is partitioned into training, validation, and test subsets, while the TLS and ULS point clouds serve as out-of-domain test sets. When analyzing the multi-platform point clouds, both the MLS and the ULS data showed positioning errors. To register them to the TLS point cloud, of which only a subset is annotated and published here, stable scene structures such as the ground, including rocks, and tree trunks had to be identified automatically, which motivated the semantic annotation into the three classes trunk (0), ground (1), and miscellaneous (2). The class distribution differs considerably between platforms. For example, tree trunks account for roughly 20% of the MLS and TLS points but only 1.7% of the ULS points, since trunks are largely occluded in the airborne acquisition. This makes the dataset particularly informative for studying cross-platform generalization. Detailed statistics on point counts and class distributions, as well as file and attribute documentation, are provided in the README accompanying the dataset. | |
Type of the data | Dataset | |
Total size of the dataset | 244787946 | |
Author | Matzke, Felix Rocco | |
Author | Eltner, Anette | |
Upload date | 2026-10-20T10:29:39Z | |
Publication date | 2026-10-20T10:29:39Z | |
Data of data creation | 2023 | |
Publication date | 2026-10-20 | |
Abstract of the dataset | Available datasets for semantic segmentation of forest point clouds mostly contain data from a single sensing system. This dataset provides LiDAR point clouds of a forest scene in the Richterschlüchte area, Saxon Switzerland, Germany, acquired during a field campaign in March 2023 using three different laser scanning systems: mobile laser scanning (MLS), uncrewed aerial vehicle laser scanning (ULS), and terrestrial laser scanning (TLS). All point clouds were manually annotated with three semantic classes, namely trunk, ground, and miscellaneous. In total, five point clouds are included: the MLS point cloud is partitioned into training, validation, and test subsets, while the TLS and ULS point clouds are provided to assess generalization across varying sensing systems. The dataset is intended for developing and testing deep learning approaches for the semantic segmentation of point clouds in complex forest environments. | |
Public reference to this page | https://opara.zih.tu-dresden.de/handle/123456789/2956 | |
Public reference to this page | https://doi.org/10.25532/OPARA-1576 | |
Publisher | Technische Universität Dresden | |
Licence | Attribution 4.0 International | en |
URI of the licence text | http://creativecommons.org/licenses/by/4.0/ | |
Specification of the discipline(s) | 3::34::315::315-02 | |
Title of the dataset | An Annotated Multi-Platform Laser Scanning Dataset (MLS, ULS, TLS) for Semantic Segmentation of Point Clouds in Complex Forest Environments | |
Software | CloudCompare v2.13.2 | |
Software | GeoSLAM Connect v1.7 | |
Software | DJI Terra | |
Software | RiSCAN PRO | |
Project abstract | Advanced Air Mobility (AAM) offers mobility around metropolitan areas in the 3rd dimension. The rapid development of small, electrically powered flight systems is promising for market readiness soon functioning as air taxi or parcel drone. Meanwhile, the fundamental technological conditions for efficient, socially acceptable operationalization of AAM still requires in-depth analysis and solutions are still in the conceptualization stage, only. To address this issue, the RTG "AirMetro" focuses on three interconnected research areas: A) Capturing the techno-economic framework, B) physical-technical modeling and experimental implementation of the AAM mobility system, and C) assessing and considering socio-technical requirements. The RTG consolidates complementary expertise in environmental science, technology, and transport safety to tackle scientifically ambitious and socially challenging research that aims to support sustainable mobility solutions beyond the operationalization of AAM. The research goal is the multi-objective modeling of a highly automated mobility system that integrates flying sensor technology, reliable and real-time capable secure communication, and mission planning of the vehicles in topologically challenging urban airspaces. The research also considers optimally located take-off/landing sites (vertiports) and their intermodal connection to the existing transport network, intelligent swarm concepts, and navigation and flight behavior under micro-climatically volatile conditions in a smart urban space. Synergies result from the intensive networking of the topics/doctoral researchers. The RTG establishes structured formats of exchange such as workshops, seminars abroad, and summer schools to foster collaboration. A postdoctoral researcher provides added value to AirMetro by scientifically linking various RTG topics aiming at concluding with an AAM concept demonstrator for an integrated mobility system. To test and validate the research, the RTG takes advantage of the brand new “Smart Mobility Lab” located in the Lusatia region, currently still under construction, which will provide high-end modern measurement and flight test equipment. All research data collected are considered highly valuable for other researchers, will be documented and archived as OpenAccess, handled by a Data Steward:ess. The RTG also provides support on an internationally excellent level through a focused qualification concept, training of professional skills, and a stay abroad under the supervision of well recognized tutors in the field. The RTG will so offer ideal conditions for an attractive, challenging career for young potentials in participating scientific disciplines and in the growing AAM industry. | |
Funding Acknowledgement | This research has been supported by the German Research Foundation (DFG, grant number 508591287) as part of RTG-AirMetro (2947/1) | |
Public project website(s) | https://tu-dresden.de/bu/verkehr/airmetro | |
Project title | RTG 2947 AirMetro - Technological & Operational Integration of highly automated Air Transport in Urban Areas |
Files
Original bundle
1 - 6 of 6
License bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- license.txt
- Size:
- 3.86 KB
- Format:
- Item-specific license agreed to upon submission
- Description:
Collections
