Back to News
Reading time:
min

Multi-Sensor UAS Hydrometry Validated at Centimeter-Level Precision in Shallow, Vegetated, and Flood-Affected River Zones

Integrated Systems
June 16, 2025

A UAV-borne system surveyed 10 km of Rönne Å river in Sweden, collecting water level, depth, and flow velocity data with an integrated suite of airborne sensors. Results were validated against ground measurements, confirming precise, contactless performance.

Background

Traditional river monitoring often relies on stationary ground-based equipment or manual sampling, which provide limited spatial coverage and are complicated to deploy in vegetated or flood-prone areas. As a result, these methods lack the resolution and flexibility needed during dynamic flow conditions.

To address this, researchers from the UAWOS project consortium, led by the Technical University of Denmark, conducted a 10 km UAS survey of the Rönne Å River in Sweden during the period from 28 August to 1 September 2023. The system combined radar altimetry, echo sounders, low-frequency ground penetrating radar (referred to as water penetrating radar, or WPR, in the terminology of the UAWOS project), and Doppler sensors to measure water level, depth, and surface velocity. The goal was to evaluate whether drone-mounted sensors could match the accuracy of ground-based methods across shallow, vegetated, and rain-affected parts of the river.

Location of the Rönne Å survey (red line) with six measured cross-sections marked as circles.

Challenges

Despite the potential of both conventional river monitoring methods and UAS-based hydrometry, field deployment revealed several technical and environmental limitations that needed to be addressed:

  • Limited spatial coverage of traditional systems. Ground-based hydrometric stations typically offer point-based measurements. In dynamic river systems, this results in data gaps that limit understanding of how water level, depth, or flow changes across space, especially during extreme weather conditions.
  • Inaccessibility in vegetated or remote areas. Many sections of the Rönne Å are challenging to access due to vegetation density, steep banks, or unstable ground. Traditional systems cannot operate in such conditions and might pose safety risks to personnel.
  • Sensor instability in moving water. Accurate sonar data requires stable, vertical positioning. Tethered sonars flown from drones are affected by current-induced tilt and drone motion. This complicates consistent depth measurements, particularly in fast-moving or uneven flow.
  • Interference in radar signal processing. Water-penetrating radar (WPR) is affected by early signal interference caused by vegetation and water surface clutter. Therefore, it’s hard to isolate bed reflections in real time or under variable flow conditions.‍
  • Airborne Doppler velocity challenges. Few precedents exist for using airborne Doppler radar in hydrometry. Accurate velocity readings require stable flight, precise angle control, and favorable water surface texture, i.e., conditions not always met during low-altitude surveys.

Solution

To test whether drone-based sensors could match the accuracy of ground equipment, researchers conducted a 10 km survey of Rönne Å using multi-sensor UAS platforms. The survey covered six cross-sections and included simultaneous ground-truth measurements for validation.

DJI Matrice 300 RTK drones were equipped with SPH Engineering’s SkyHub and True Terrain Following system for automated low-altitude flight. The drones carried:

  • Radar altimeters: Geolux LX-80 for measuring water surface elevation.
  • Echo sounder: ECT400S for depth mapping.
  • Water-penetrating radars (WPR): Zond Aero LF for contactless bathymetry through vegetation.
  • Surface velocity sensors: Geolux RSS-2-300W Doppler radar and GoPro Hero 8 Black for video-based PIV (particle image velocimetry).‍
  • LiDAR and photogrammetry tools: DJI Zenmuse L1 for terrain modeling.

Each sensor stream was processed in three steps: raw data logging, georeferencing and synchronization, and spatial projection to river cross-sections. Rainfall on 31 August created flood-like conditions, allowing validation under hydrologic stress. Ground-based systems were co-deployed to enable direct, point-by-point performance comparison.

Field deployment of M300 drone with ECT400S echo sounder (left) and Zond Aero LF WPR (right)

Results

Water Surface Elevation (WSE)

Radar altimeters mounted on DJI Matrice 300 RTK measured water surface elevation with vertical precision under 3 cm. Comparisons with GNSS benchmarks and earlier radar sensors confirmed strong alignment and consistency.

Representative comparisons between airborne and ground-based WSE data: (a) Flight paths, in-situ WSE points, and river centerline. (b–e) WSE comparisons between ground truth and four LX80 configurations.
More about Contactless Airborne Hydrometry

Bathymetry

Echo sounder and water-penetrating radar (WPR) systems provided complementary depth measurements. The ECT400S offered stable sonar-based profiles with deviations under 10 cm. In contrast, the Zond Aero LF GPR (WPR) system successfully penetrated vegetation and identified bed contours that echo sounders could not detect.

All depth measurements were validated against ADCP readings, showing WPR accuracy within ±8 cm.

Comparison of drone- and ground-based bathymetry across six cross-sections. Includes M300 with ECT400s (magenta), Hydrohex with ECT400 (yellow), M600 with Deeper CHIRP+ (cyan), in-situ ADCP reference (black stars), and drone-based lidar (green triangles). Blue lines represent water surface elevation.

Flow Velocity

Surface flow velocity was captured using the Geolux RSS-2-300W Doppler radar and analyzed alongside video-based particle image velocimetry (PIV) from a waterproof action camera (GoPro Hero 8 Black). The Doppler system maintained a resolution of 7.3 mm/s and achieved deviations below 10 cm/s when benchmarked against OTT MFPro instruments. PIV results showed variation, particularly in vegetated cross-section XS1, but aligned well at other sites.

Velocity comparisons across five cross-sections using Doppler radar (triangles), OTT MFPro (black stars), and PIV (red points, within 2.75 m of the tagline). In XS5, Doppler results are shown for two altitudes—2.1 m and 4.1 m—before and after rainfall.

Conclusion

The Rönne Å River survey demonstrated that a multisensor contactless drone-based system can accurately measure water level, depth, and flow speed, even in shallow, vegetated, and flood-affected zones. Supported by SPH Engineering’s integrated sensors, SkyHub and True Terrain Following, the system provided precise, low-altitude survey results that closely matched ground-based references.

The publicly shared dataset supports broader research and shows the potential of drone-based hydrometry as a practical alternative to conventional methods.

Read the full report

Related Publications

About
Download UgCS