Skip to content

SpectralBridge logo

SpectralBridge

SpectralBridge (formerly cross-sensor-cal) translates reflectance among NEON AOP, uncrewed aerial system (e.g., MicaSense), and Landsat observations to place disparate datasets into a common Landsat-referenced frame. It is designed for notebook-first scientists who need reliable, reproducible harmonization to bridge fine-scale ecological measurements with the long-term continental Landsat record. The workflow is pipeline-based rather than a single correction, and it produces artifacts that can be inspected, reused, and resumed.

Upgrading from older versions? Imports under cross_sensor_cal and legacy cscal-* CLI aliases still function, but new examples use the spectralbridge namespace and spectralbridge-* commands.

What SpectralBridge does

  • Normalizes directional reflectance through topographic and BRDF adjustments, then resamples spectra into Landsat and other target bandpasses for direct comparison.
  • Bridges across ecological scales by aligning field, UAS, and airborne measurements with the Landsat time series, enabling cross-scale analyses and model transfer.
  • Generates harmonized, provenance-rich spectral libraries alongside imagery so downstream modeling can start from analysis-ready data.
  • Emits restart-safe outputs (ENVI, Parquet sidecars, merged pixel tables) plus QA panels that document every run.

Why this approach is trustworthy

  • Physics-informed normalization (topographic + BRDF) precedes empirical calibration to keep results grounded in surface reflectance rather than image brightness alone.
  • Cross-sensor calibration is referenced to Landsat NBAR, providing an anchored target for spectral translation.
  • Every run produces QA PNG/PDF panels and JSON summaries so users can interrogate performance and document decisions.
  • The pipeline is deterministic and restartable: intermediate artifacts are written to disk, making runs auditable and recoverable.

Connection to the scientific literature

This software implements the workflow described in the Remote Sensing of Environment manuscript, Bridging Scales for Macrosystems Ecology: Harmonizing Western US Plant Functional Types Spectral Data from Drones and NEON Airborne Imagery to Landsat Observations. It operationalizes that study's approach to aligning UAS and NEON hyperspectral data with Landsat, providing an open, inspectable implementation for reproducible science.

Who should use this

  • Notebook-first scientists analyzing spectral libraries and reflectance products.
  • Ecologists and remote sensing researchers needing to connect plot-level or flightline observations to Landsat's long-term record.
  • Advanced users and developers integrating the pipeline into larger workflows or extending it to additional sensors.

Next steps