Mi et al. (2026) A UAV-Based Dual-Spectroradiometer Method for Hyperspectral Reflectance Measurement

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-04-05
• Authors: Haoheng Mi, Yu Zhang, Hong Guan, Kang Jiang, Yongchao Zhao
• DOI: 10.3390/rs18071093

Research Groups

• Key Laboratory of Target Cognition and Application Technology (TCAT), Chinese Academy of Sciences, Beijing 100190, China
• Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
• School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing 100049, China
• UTAN Technology Co., Ltd., Hangzhou 310012, China

Short Summary

This study developed and validated a UAV-based dual-spectroradiometer system for hyperspectral surface reflectance measurement under natural illumination. The system achieved accurate irradiance and reflectance measurements, with RMSE below 0.01 and Spectral Angle Mapper (SAM) values below 3.5 across the 400–900 nm spectral range, demonstrating a practical and ground-independent approach for quantitative hyperspectral reflectance acquisition.

Objective

• To develop and validate a UAV-based spectroradiometric system for surface reflectance retrieval under natural illumination conditions using non-imaging hyperspectral sensors, enabling reflectance retrieval through a radiance–irradiance ratio framework without relying on ground calibration targets or radiative transfer model inversion.

Study Configuration

• Spatial Scale: Agricultural plots (Field 1–Field 4) at Jiusan Farm, Heihe, Heilongjiang Province, northeastern China (48°58′41″N, 125°16′28″E). UAV flight altitudes were approximately 30–40 meters above ground level (AGL).
• Temporal Scale: Two flight missions conducted on the same day (Flight 1: approximately 10:34–10:46 local time; Flight 2: approximately 12:24–12:34 local time). Data acquisition interval was 100 milliseconds.

Methodology and Data

• Models used: Radiance–Irradiance Ratio formulation (for reflectance retrieval), MODTRAN 4.3 (for atmospheric effects simulation).
• Data sources:
  • UAV-based system: Custom-developed dual-spectroradiometer system integrating two non-imaging hyperspectral spectrometers (Ocean Optics USB2000+ VIS–NIR, 400–900 nm spectral range, <1 nm spectral resolution, >200:1 signal-to-noise ratio, ~10 Hz sampling rate) mounted on a quadcopter UAV. The spectrometers were on high-precision stabilized gimbals (pitch and roll standard deviations below 0.5°).
  • Ground-based validation: Two ASD FieldSpec 4 spectroradiometers (350–2500 nm spectral range) for continuous downwelling irradiance monitoring and in situ upwelling surface radiance measurements.
  • Auxiliary data: Onboard GPS modules for time synchronization and positioning, Inertial Measurement Unit (IMU) for gimbal stabilization, CE318 sun photometer for atmospheric properties.

Main Results

• The UAV-based dual-spectroradiometer system successfully acquired stable downwelling irradiance and upwelling surface radiance measurements during flight.
• UAV-based irradiance measurements showed strong consistency with ground-based ASD reference observations, with residuals generally confined within 0.05 across the 400–900 nm spectral range, even during UAV turning maneuvers.
• UAV-derived surface reflectance demonstrated good agreement with ground-based spectroradiometric measurements, yielding Root Mean Square Error (RMSE) values generally below 0.009 and Spectral Angle Mapper (SAM) values below 3.5 degrees across the 400–900 nm spectral range.
• Atmospheric effects were found to be minor under the study's low flight altitudes (30–40 meters) and favorable clear-sky conditions, with atmospheric transmittance remaining above 0.99 and path radiance contributing less than 1% of the measured signal for most wavelengths.
• Consistent spectral deviations were observed, where UAV-measured irradiance and derived reflectance were slightly higher in the 520–580 nm range and slightly lower in the 600–710 nm range compared to ground measurements, attributed to differences in the surrounding illumination environment.

Contributions

• Developed a practical and reliable UAV-based dual-spectroradiometer method for quantitative hyperspectral surface reflectance retrieval under natural illumination, offering a ground-independent approach without explicit atmospheric correction under specific conditions.
• Enhanced spatiotemporal consistency of UAV-based reflectance measurements through a system-level design incorporating actively stabilized gimbals for both upward-looking irradiance and downward-looking radiance sensors.
• Provided rigorous field validation of the UAV-derived reflectance against synchronized ground-based dual-spectroradiometer measurements, demonstrating high accuracy and spectral fidelity.
• Offered practical insights and guidance for UAV flight planning and radiometric system design, particularly regarding the impact of platform maneuvers and environmental geometry on irradiance measurements and subsequent reflectance retrieval.

Funding

• Key Laboratory of Target Cognition and Application Technology (grant number E4M9080306).

Citation

@article{Mi2026UAVBased,
  author = {Mi, Haoheng and Zhang, Yu and Guan, Hong and Jiang, Kang and Zhao, Yongchao},
  title = {A UAV-Based Dual-Spectroradiometer Method for Hyperspectral Reflectance Measurement},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18071093},
  url = {https://doi.org/10.3390/rs18071093}
}