Skip to content

Choose a vignette

SpectralBridge has three separate pipelines. Choose the one that matches the data you have; the bulk pipeline analyzes completed flightline products and does not run NEON or drone processing for you.

Pipeline Start here Runnable notebook What it does
NEON flightlines Full NEON vignette NEON notebook Download or read NEON HDF5; correct, convolve, extract, and review QA for individual flightlines.
Drone imagery Drone vignette Drone notebook Read local TIFF/HDF5; correct native MicaSense; optionally apply packaged affine coefficients, extract, and review QA. No convolution.
Bulk analysis Bulk vignette Local bulk notebook Read already completed, immutable flightline products; fit population relationships and interpret compact outputs.
Remote drone production Production vignette CyVerse production notebook Inventory remote ExportPackages, process one flight at a time, enforce completeness, run bulk/reporting, and optionally upload verified compact results.

The bulk coefficient evidence comes from synthetic matched products of the same corrected NEON source. It is diagnostic, not universal empirical sensor calibration. Drone use of those coefficients remains an explicit opt-in with band-level cautions and a corrected-value scale check.

Start with your situation

I want to… Vignette
Run everything for a NEON flightline Run the full pipeline
Continue after a stopped or partial run Carry On My Wayward Son
Catalog many completed runs and analyze a sensor population Build a bulk cross-run analysis
Process CyVerse drone ExportPackages and produce a bulk closeout package Remote drone production
Work on one part of the workflow Choose a module below
Open a runnable Jupyter notebook Runnable notebook vignettes
Look up exact arguments, filenames, or algorithms Technical reference

Module vignettes

The numbered modules follow the order of the main NEON workflow. Drone and polygon workflows branch from that sequence where noted.

  1. Acquire NEON data — download or reuse the source HDF5.
  2. Correct NEON reflectance — create raw and BRDF/topography-corrected ENVI products.
  3. Harmonize target sensors — translate corrected spectra into configured target-sensor bandspaces.
  4. Build analysis tables — work with per-product and merged Parquet outputs for an individual run.
  5. Review QA outputs — render and interpret the PNG and JSON audit artifacts.
  6. Process drone imagery — run the separate local-HDF5 drone workflow.
  7. Extract polygon spectra — build polygon-indexed spectral libraries from completed products.
  8. Build a bulk cross-run analysis — catalog canonical flightlines, derive narrow observations from persisted target-sensor ENVI, and fit pooled, balanced, and held-out-site synthetic translations without modifying individual runs. Prebuilt merged Parquets remain supported.

How to use these pages

Every vignette answers the same four questions:

  1. When should I use this module?
  2. What is the smallest runnable example?
  3. What files prove that it worked?
  4. Where are the deeper technical details?

The vignettes teach workflow. The technical reference defines contracts and implementation details.

Prefer code you can edit interactively? Every module also has one corresponding downloadable notebook, including a carefully isolated template for adding a custom correction between the standard topo/BRDF and convolution stages.