Extremely dense LiDAR data (Light Detection and Ranging) in combination with very high resolution digital ortho-images recorded using a helicopter-based platform offer new possibilities for feature extraction. Here the authors demonstrate the use of such data for the extraction of railroad infrastructure objects. Thus, the processes of data capturing, the pre-processing as well as different methods for the extraction of railroad infrastructure objects, such as rail tracks/centre lines, railway poles, cantilevers and catenary are introduced. In this study this kind of data were analysed in order to extract railway infrastructure equipment and its surrounding objects. The main focus is on studies for a fully automated processing in consideration of profitability aspects compared to traditional surveying methods. A further objective is the development of methods for the object extraction of railway infrastructure from combined helicopter-based extremely dense laser scanner measurements and very high resolution digital ortho-images. The developed methods are aimed at the generation of high-precision input data to create and extend a railroad infrastructure database. The required high density of laser points can only be achieved with high quality equipment and a slowly flying aircraft. For being able to detect small and thin elements such as rail tracks the relative (and absolute) accuracy of the laser points must be at a few centimetre level. To meet this goal the positioning of the aircraft (and the subsequent georeferencing) has to be at the same accuracy level. The area under investigation is a 36 km railroad section between Korneuburg and Hollabrunn, north of Vienna, Austria. The corridor contains single- and double-track sections as well as some small railroad station areas and is property of the Austrian Federal Railway (ÖBB-Infrastruktur Bau AG), who is also the initiator of this study.
Automated object extraction from LiDAR and ortho-image data for railroad databases
Automatische Objektextraktion aus LiDAR- und Ortho-Bilddaten für Eisenbahndatenbanken
2008
6 Seiten, 3 Bilder, 1 Tabelle, 4 Quellen
Conference paper
English
/LiDAR AUTOMATED OBJECT ANNOTATION USING FUSED CAMERA/LiDAR DATA POINTS
European Patent Office | 2025
/LiDAR AUTOMATED OBJECT ANNOTATION USING FUSED CAMERA/LiDAR DATA POINTS
European Patent Office | 2025
|/LiDAR AUTOMATED OBJECT ANNOTATION USING FUSED CAMERA/LiDAR DATA POINTS
European Patent Office | 2023
|Automated Highway Sign Extraction Using Lidar Data
Transportation Research Record | 2017
|Automated object annotation using fused camera/LiDAR data points
European Patent Office | 2024
|