Integration of LiDAR Technology in Bicycles for Updating 3D Models of Urban Infrastructure.

Abstract

This study assesses the feasibility of a low-cost, bicycle-mounted LiDAR SLAM and spherical camera system for efficient urban data acquisition, addressing the high costs and accessibility limitations of conventional mobile and airborne LiDAR.

Conventional methods struggle with accurate inventory of urban furniture and tree canopies due to vehicle accessibility restrictions (mobile LiDAR) and data density limitations (airborne LiDAR). The Omaha, Nebraska study comprised field data acquisition (planning, execution, backup, quality control) and data analysis (trajectory processing, point cloud generation, comparison with public data). A lightweight, low-cost LiDAR SLAM 100 sensor provided optimal point density for semi-automated urban furniture identification, complemented by a cost-effective Ricoh Theta Z camera for 360° imagery and street view generation. Results showed a 0.045-meter average vertical difference from public data, with a standard deviation of 0.01 cm, validating the method’s accuracy. Semi-automated 3D line extraction and point cloud classification enhanced efficient urban element detection. The system’s speed and processing capabilities make it a viable alternative for detailed mapping in pedestrian areas and under bridges, supplementing traditional methods, particularly for smaller update areas. This approach facilitates digital twin creation and efficient public space management at a significantly lower cost than conventional techniques.

Introduction

Accurate and current urban mapping is essential for effective city planning and management. Traditional methods, such as airborne and vehicle-mounted LiDAR, offer broad coverage but lack the detail needed for modern urban environments, especially in pedestrian areas, under bridges, or in areas with dense vegetation. Airborne LiDAR lacks the resolution for fine details; UAV LiDAR struggles under tree cover and lacks point density for pedestrian areas; and vehicle-mounted LiDAR is limited to roads, excluding many important urban spaces. These limitations necessitate innovative and cost-effective solutions.

This research proposes a bicycle-mounted LiDAR system as a novel approach to address these shortcomings. This system provides a pedestrian perspective, offering superior detail for applications such as pavement assessment, signage optimization, route planning, and precise urban inventory updates. The bicycle’s maneuverability allows access to restricted areas with minimal traffic disruption, and its speed optimizes data collection. Integration with 360° imagery enhances map creation and navigation systems, similar to Google Street View. This methodology is particularly effective in challenging environments where traditional methods fail.

Omaha, Nebraska, serves as a relevant case study. Its commitment to innovative mobility solutions and rapid urban development necessitates ongoing cartographic updates. This research aligns with the Omaha Metropolitan Area Planning Agency’s 2025 Unified Plan, supporting efforts to improve regulations, enhance cycling and pedestrian infrastructure, and expand GIS datasets. This study offers a detailed and cost-effective method for acquiring high-resolution urban data, overcoming limitations of existing techniques.

While existing research demonstrates advancements in mobile LiDAR (Feng et al., 2018; Castaneda & Castaneda, 2019), prior work primarily focuses on vehicle and drone platforms. This study addresses the gap in research on bicycle-mounted LiDAR for detailed urban mapping, particularly in pedestrian areas. This innovative approach aims to improve accessibility, reduce costs, and enhance the precision of urban data acquisition. We will evaluate the system’s performance, precision, and processing efficiency, assessing its capacity to generate detailed 3D models with attributes. This research will contribute to the development of sustainable and accessible urban mapping solutions, ultimately improving urban planning and management.

Methodology:

This study utilized a novel, cost-effective bicycle-mounted LiDAR system (LiDAR SLAM 100, $25,000 USD including software; Ricoh Theta Z camera, $1,000 USD) for high-resolution urban data acquisition, overcoming limitations of traditional airborne and vehicle-based LiDAR. The system’s lightweight design (1.6 kg) and high accuracy (2cm relative, 5cm absolute) were key considerations.  The Ricoh Theta Z camera provided 23MP images and 4K video for detailed 360° visual context and streamlined Google Street View integration.

Mounted on a mountain bike with a rear rack, the system was secured to minimize vibrations. Data acquisition occurred on Omaha’s Keystone Trail, selected for minimal traffic and the presence of various urban features. Data collection occurred under controlled environmental conditions (3°C, clear skies, moderate winds).

Data processing involved SLAM GO POST for trajectory correction and georeferencing (using the NGS NEOM CORS permanent station), TOPODOT for 3D modeling and point cloud classification, and Global Mapper and QGIS for DTM generation and comparison with public data. Data validation compared bicycle-acquired data with public LiDAR data, focusing on paved, vegetation-free areas to assess vertical differences (mean, standard deviation, completeness) and horizontal accuracy via cross-sectional analysis.  This rigorous methodology ensured high-quality data for evaluating the system’s efficacy in urban mapping.

Results and discussion:

This study demonstrates the effectiveness of a bicycle-mounted LiDAR system for high-resolution urban mapping, achieving a data acquisition speed of 15 km/h on Omaha’s Keystone Trail without compromising data quality. The system’s lightweight design and efficient mounting (5 minutes for setup and takedown, facilitated by access to the NEOM CORS station) allowed for seamless integration within the Omaha metropolitan area. No significant data acquisition challenges were encountered.

Data validation against publicly available LiDAR data yielded high accuracy: a 0.045-meter (4.53 cm) average difference and 0.013-meter (1.34 cm) standard deviation, indicating strong consistency. The lower completeness (8.77%) is attributed to the significant density difference between the bicycle-acquired data (6500 pts/m²) and the public data (10 pts/m²). No systematic errors or significant deviation patterns affecting data quality were identified.

The system facilitated semi-automated extraction of various urban features: 3D lines representing road edges, curbs, fences, and other delineations; utility poles and power lines with height and type attributes; traffic signs with height and type attributes; trees with diameter estimates; utility boxes and manholes; and bridges with height and dimensional estimates (Figure 6: Perspective and Details Captured in Bike Data). Linear features were segmented and classified using TOPODOT algorithms; poles and signs were identified via semi-automated detection with manual validation in MicroStation Open Roads.

Compared to vehicle-mounted LiDAR, the bicycle system shows a reduced horizontal range (75 meters vs. 500 meters) due to sensor height and range limitations. However, data density is comparable, enabling detailed 3D modeling. The bicycle-LiDAR method offers several advantages: significantly lower cost than vehicle or drone-based systems; access to pedestrian and restricted areas; reduced vehicular interference and associated permitting complexities; enhanced portability and flexibility; and the unique ability to capture detailed data under bridges. However, limitations exist: reduced horizontal range requiring multiple passes for wider areas; confinement to paved surfaces and designated cycle paths for operator safety; and potential limitations in areas with dense vegetation or buildings that may obstruct GPS signals (although not encountered in this study).

A cost comparison with other mobile mapping technologies (Khanal et al., 2020) shows significant cost advantages for the bicycle-based approach. While not directly comparable due to differing methodologies and scopes, the cited study (2016 data) suggests a substantially higher cost per kilometer for traditional topographic surveys, airborne LiDAR (both large-area and single-project scales), and terrestrial mobile LiDAR, while UAV photogrammetry shows lower costs (Topographic Survey: $7521.52/km; Airborne LiDAR (large area): $421.77/km; Airborne LiDAR (single project): $6420.45/km; Terrestrial Mobile LiDAR: $4590.19/km; UAV Photogrammetry: $1300/km). Our estimated cost for bicycle-based LiDAR for urban furniture mapping and 3D modeling in 75-meter-wide corridors is $95/km. While the linear cost is lower, the system’s limited range restricts its competitiveness with large-area mobile mapping systems. This cost-effectiveness, combined with its unique accessibility and data quality, positions the bicycle-mounted LiDAR system as a valuable addition to the urban mapping toolkit, especially for detailed surveys in specific areas.

Conclusions:

This study successfully demonstrates the feasibility and cost-effectiveness of bicycle-mounted LiDAR for high-precision urban data acquisition. The method achieved a 15 km/h data acquisition speed with minimal vibration impact on sensor precision, requiring only 5 minutes for setup and takedown. Data validation against public LiDAR data showed a 4.53 cm average vertical difference and 1.34 cm standard deviation, confirming high accuracy despite a lower completeness due to differing data densities. The system enabled semi-automated extraction of detailed 3D models of various urban features, including road edges, poles, signs, trees, and bridges, enhancing efficiency compared to traditional methods.

This approach shows significant potential for modernizing urban mapping, particularly in areas with limited vehicular access (e.g., cycle paths, secondary streets, pedestrian zones, under bridges). The rapid and accurate extraction of urban elements facilitates efficient public space management. An efficiency analysis suggests that Omaha’s main cycle paths could be updated in 21 days at an estimated cost of $12,920, significantly reducing time and operational costs compared to traditional methods. The cost-benefit analysis shows superiority to traditional methods and, while not a replacement, provides a valuable complement to airborne and mobile LiDAR systems, filling data gaps at a lower cost. The approach enhances urban planning by providing rapid access to comprehensive data on specific areas, improving response times for information acquisition. The system’s versatility makes it suitable for municipalities and smaller communities seeking detailed mapping without significant investment.

Further considerations include access to control networks, validation points, and designated cycle paths.

Future improvements include evaluating data acquisition in areas with tall buildings to assess potential signal obstruction effects, exploring alternative small-scale mobile platforms (e.g., electric vehicles) to enhance coverage and efficiency, and investigating automated LiDAR data extraction techniques to optimize post-processing and reduce analysis time. This research advances the use of alternative mobility platforms for spatial data acquisition, offering cost-effective and high-impact solutions for urban mapping.

References:

• Castaneda, M., & Castaneda, D. (2019). Actualización Catastral en Bogotá informacion Planimetrica y Altimetrica usando Drone con LIDAR. Bogota: Universidad Distrital Francisco Jose de Caldas.
• Feng, F., Bao, S., Hampshire, R., & Delp, M. (junio de 2018). Drivers overtaking bicyclists—An examination using naturalistic driving data. Accident Analysis & Prevention, págs. 98-109.
• Khanal, M., Hasan, M., Sterbentz, N., Johnson, R., & Weatherly, J. (2020). Accuracy Comparison of Aerial Lidar, Mobile-Terrestrial Lidar, and UAV Photogrammetric Capture Data Elevations over Different Terrain Types. Infrastructures 5, no. 8: 65, 65.
• Metropolitan Area Planning Agency. (2024). Unified Planning Work. Omaha: Mapa .
• Topodot. (30 de 1 de 2025). https://blog.topodot.com/. Obtenido de Establishing Requirements, Extracting Metrics, and Evaluating Quality of LiDAR Data: https://blog.topodot.com/establishing-requirements-extracting-metrics-and-evaluating-quality-of-lidar-data/