BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//talks.staging.osgeo.org//foss4g-europe-2025//talk//3EPQL E BEGIN:VTIMEZONE TZID:CET BEGIN:STANDARD DTSTART:20001029T040000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET TZOFFSETFROM:+0200 TZOFFSETTO:+0100 END:STANDARD BEGIN:DAYLIGHT DTSTART:20000326T030000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST TZOFFSETFROM:+0100 TZOFFSETTO:+0200 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:pretalx-foss4g-europe-2025-3EPQLE@talks.staging.osgeo.org DTSTART;TZID=CET:20250717T113000 DTEND;TZID=CET:20250717T120000 DESCRIPTION:Autonomous driving approaches require simulation environments t hat accurately converge real-world conditions. We can utilize 3D reconstru ction methodologies to achieve this. In this paper\, we propose a lightwei ght 3D reconstruction methodology by using the Geospatial Data Abstraction Library (GDAL) and existing HD maps in the OpenDRIVE data format. By leve raging these datasets\, we aim to improve the efficiency and accessibility of 3D scene reconstruction for autonomous driving applications. Additiona lly\, we aim to provide a low-cost solution to address the annotation bott leneck in point-wise labeling for the computer vision domain with the cons tructed 3D models.Scene understanding is crucial for autonomous driving\, and most approaches rely on online sensor measurements to extract meaningf ul information from self-driving cars' surroundings. Visual sensors like c ameras and Light Detection and Ranging (LiDAR) are used a lot in autonomou s driving for perception tasks. These observations\, along with measuremen ts from Inertial Measurement Units (IMU) and Global Navigation Satellite S ystems (GNSS)\, can help us understand scenes better. However\, the commun ity realized that environmental factors such as weather conditions and ill umination can easily affect those measurements. Then high-precision geospa tial data became a lifesaver for automated driving. Last decade\, the auto motive domain has had tremendous interest in high-definition (HD) maps. HD maps are essential and complementary to achieving accurate navigation and localization for autonomous driving. The performance of current perceptio n sensors is limited by their surroundings. This can make it hard for self -driving cars to figure out where they are\, especially when they are near the edges of their surroundings\, which can make the experience of drivin g unsafe. Problematic situations include GNSS receivers being affected by multipath effects\, LiDAR systems not being able to scan beyond a certain range\, and camera image processing algorithms not being able to get clear images that can be used. On the other hand\, visual sensors (camera and L iDAR) need pre-annotated and trained datasets to predict and detect object s that are visible in the scene. In contrast to those conditions\, HD maps can provide static lane-level information\, which allows for an informati ve baseline about the surroundings in any case.The most common HD map form ats include Navigation Data Format (NDS)\, OpenDRIVE\, and Lanelet2. Howev er\, differences in road geometry definitions and data structures make cro ss-format compatibility difficult. Each format has distinct ways of storin g and structuring road elements\, creating obstacles for seamless data int egration across platforms. OpenDRIVE\, an industry-standard format develop ed by the Association for Standardization of Automation and Measuring Syst ems (ASAM)\, provides a structured way to describe road networks in lane-l evel detail.In this work\, we will investigate OpenDRIVE's ability to crea te 3D shapes and explore realistic convergence strategies for autonomous d riving simulations. Using the GDAL XODR driver—available since 2024—we will generate 3D geometries as OGC Simple Features for selected road area s. This driver allows the creation of Triangular Irregular Networks (TIN) from driving surfaces and 3D road infrastructure elements\, which we will use to generate synthetic point clouds. By leveraging these synthetic poin t clouds\, we can systematically evaluate how well vector-based models app roximate real-world environments. This enables a direct comparison between vector-based 3D modeling and real-world LiDAR data/point clouds.Point clo uds have been widely used for scene understanding in autonomous driving\, as they provide 3D coordinates and intensity values for the environment. H owever\, large-scale 3D modeling is computationally expensive and requires efficient data processing techniques. Annotating these datasets manually is also time-consuming and labor-intensive\, making semantic information e xtraction difficult. The lack of automation in labeling further exacerbate s these challenges\, slowing down the development of advanced perception m odels. Addressing these limitations is essential to improving HD map appli cations and integrating them into broader geospatial workflows.We will use the Iterative Closest Point (ICP) algorithm to make sure that synthetic a nd real-world data are more closely aligned. This will reduce errors and a llow for accurate shape reconstruction. The ability to refine and align sy nthetic models with real-world measurements is crucial for ensuring high-f idelity simulations. Additionally\, we will use Nearest Neighbor Search an d Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to identify corresponding points and fill missing HD map elements. By intelli gently reconstructing missing data\, we can enhance the completeness of HD maps and make them more reliable for self-driving applications. This will enhance data completeness and improve overall map reliability for self-dr iving applications.By integrating these methodologies\, we aim to bridge t he gap between vector-based HD maps and real-world point cloud data. This will enable a more seamless fusion of geospatial information across differ ent domains\, improving both data usability and accuracy. Our method aims to make a 3D reconstruction pipeline that is more accurate and faster. Thi s will make it easier to simulate self-driving cars and help validate and improve HD maps. Ultimately\, by enhancing the accuracy and efficiency of 3D modeling techniques\, our approach contributes to safer and more effect ive autonomous driving systems. DTSTAMP:20260527T015053Z LOCATION:PA01 (Quarticle) SUMMARY:Integration of HD Maps and Point Clouds: An Efficient 3D Reconstruc tion Framework for Autonomous Driving Applications - Gülsen Bardak URL:https://talks.staging.osgeo.org/foss4g-europe-2025/talk/3EPQLE/ END:VEVENT END:VCALENDAR