Drone-based road survey

Road inspection is a demanding task that requires time, equipment and repeated effort, while conventional surveys can be costly and disruptive for traffic. Drone-based sensing offers a way to reduce this effort and make road monitoring more flexible, repeatable and frequent.

We have developed and experimentally validated a method for measuring road geometry and assessing vehicle-relevant road parameters using data from drone-mounted sensors. This approach enables accurate capture of road configuration, precise geometry, surface defects, wear indicators and other parameters required for digital road-condition mapping.

  • Faster road-condition assessment compared with conventional inspection methods
  • Autonomous and repeatable monitoring of selected road sections
  • Frequent inspections without blocking or significantly disturbing traffic
  • Accurate capture of road geometry, defects, wear and surface parameters
Drone scanning a curved road corridor

Wide-area sensing from above

Drone-based road corridor observation from above

A road corridor can be observed as a complete scene from above. This perspective makes it possible to capture the lane area, road edges and surrounding objects in one measurement pass.

Compared with vehicle-mounted sensors, drone-mounted sensors are less limited by the low viewing position. Vehicles, roadside objects or temporary obstacles can restrict the field of view of sensors installed at traffic level and hide relevant parts of the road scene.

From an elevated position, the drone provides a wider effective field of view. This allows the system to measure more road parameters at once, including road configuration, precise geometry, lane boundaries, visible defects, wear indicators and objects located near the carriageway.

The top-view perspective is especially useful for structured road-condition mapping, because it shows not only where an object is located, but also how it relates to the driving path and to the surrounding road environment.

Flight speed up to 2 m/s
Survey width up to 3 lanes at once
Flight altitude about 10 m above the road
Road length per battery about 4 km
Top view of a road scene with vehicles, obstacles and roadside objects
Road scene from a vehicle-level perspective with limited visibility

From drone measurements to road parameters

Dense surface reconstruction The drone collects sensor data from above and creates a dense surface mesh of the inspected road section.

Detailed road analysis The reconstructed surface makes it possible to analyse road geometry and wear in detail, including rutting, slope, local depressions and other surface irregularities.

Geo-referenced parameters The extracted parameters are linked to their position on the road, allowing defects, geometry changes and wear indicators to be documented in a coordinate-based road-condition map.

Processing of drone-based road measurements into texture, elevation and road parameter layers

Surface profile, texture and defects

Measured road surface profiles and texture indicators

Road condition is described not only by visible defects, but also by the measured surface structure. Profile and texture information make it possible to detect local changes that are relevant for maintenance and vehicle behaviour.

The measured data can be used to analyse longitudinal and transverse evenness, roughness, local height changes and surface irregularities. This helps to distinguish smooth asphalt, worn pavement, rough surfaces and unpaved sections.

Together with the coordinate-based road map, these indicators support repeatable inspection, documentation of road wear and comparison of the same road section over time.

```

Multi-layer road-condition map

The drone survey creates a multi-layer map of the road. Visual, geometric and profile information are combined and linked to coordinates, so that defects, road parameters and condition indicators can be located precisely.

Different sensor layers describe different aspects of the same road section. A high-resolution camera supports visual inspection and the creation of road textures. Scanning LiDAR is used to reconstruct the road profile and 3D road geometry, including slope and gradient. Single-point LiDAR can measure selected lane lines and support the analysis of road microstructure, profile and wear.

Map layer Measured information
Visual layer Road texture, visible defects, objects and obstacles.
Geometry layer Road edges, lane position, slope, gradient and 3D road shape.
Profile layer Longitudinal and transverse evenness, local height changes and roughness.
Condition layer Wear indicators, surface condition and estimation of skid-resistance-related parameters.
Coordinate layer Geo-referenced road-condition information for documentation and maintenance planning.
Multi-layer road-condition map with texture, geometry and profile layers

The road is represented as several coordinated layers: visual texture, geometric reconstruction and measured road-profile information.

Why use a UAV for road monitoring?

Autonomy

Repeated inspections can be carried out with predefined missions. With docking or charging infrastructure, regular autonomous operation becomes possible.

Safety

The drone does not need to block the road during measurement and can observe a larger area than a vehicle equipped only for direct surface measurement.

Modularity

The sensor payload can be selected according to the required output: geometry, surface profile, visual inspection, defect detection or friction-related indicators.

Digital infrastructure

The result is a geo-referenced road-state map that can support maintenance decisions and vehicle-to-infrastructure information services.