speclab

Python spectroscopy processing library for thermal infrared emissivity calibration, FTIR transmission/reflectance calibration, and VSWIR reflectance analysis. Ports and extends algorithms from the DaVinci spectroscopy environment.

Features

Emission calibration

• emcal — full lab emission calibration pipeline (blackbody IRF, NEM / hullfit / alpha / MMD emissivity retrieval, optional noise-free IRF smoothing, water-vapour correction)
• dehyd — water-vapour correction applied during or after emcal

Emissivity retrieval methods

• Normalized Emissivity Method (nem) — iterative BT peak anchoring
• Convex-hull Planck mixture (hullfit / hullfit_linear) — strict upper-bound enforcement
• Alpha Residuals (alpha) — mean-BT reference with max-emissivity rescaling
• Maximum–Minimum Difference (mmd) — simple contrast-based baseline

FTIR transmission / reflectance calibration

• tracal — transmission calibration from an AutomateFTIR measurement folder; pairs each sample with its closest-in-time background and blank
• refcal — reflectance calibration; same pipeline as tracal, reflectance semantics

VSWIR reflectance analysis

• remove_continuum — convex-hull continuum removal
• band_parameters — per-feature band depth, FWHM, area, asymmetry
• smooth_spectrum — Savitzky-Golay, boxcar, Gaussian smoothing
• detect_bands — automatic absorption feature detection with preset matching

Spectral analysis

• sma — spectral mixture analysis (NNLS, grouped endmembers, slope endmember, cumulative overlay plots)
• merge — combine emcal outputs vertically (same range, new samples) or horizontally (same samples, extended spectral range with DC-offset alignment)
• match — merge two individual spectra covering overlapping but different spectral ranges

Data I/O

• readDVhdf / saveDVhdf — DaVinci-format HDF5 read/write
• dv_to_album — normalise any DV HDF5 layout to the per-spectrum album dict
• loadReflectanceCSV / saveReflectanceCSV — wide-format VSWIR reflectance CSV
• loadASD — ASD ViewSpecPro tab-separated text export
• convert_speclib — command-line tool to convert spectral libraries: DV HDF5 → SpectralViewer format, USGS splib07 → speclab HDF5, CRISM spectral library → speclab HDF5

Instrument presets for NAU, ASU, SwRI lab spectrometers and TES / mTES satellite instruments

GUI tools

Command / Script	Module	Platform	Purpose
speclib-viewer	SpeclibViewer	all	Browse and build spectral libraries (LWIR / VSWIR)
emission-processor	EmissionLWIR	all	Interactive emcal / SMA results viewer
reflectance-vswir	ReflectanceVSWIR	all	VSWIR reflectance viewer and band analysis
AutomateFTIR.pyw	—	Windows only	Automated FTIR data collection (OMNIC DDE + Keithley 2700)

Installation

Using conda / mamba (recommended)

conda create -n speclab python=3.11
conda activate speclab
pip install -e .

Using venv

python3 -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -e .

Both approaches install all core dependencies listed in pyproject.toml and make the speclib-viewer, emission-processor, and reflectance-vswir entry-point commands available inside the environment.

For the AutomateFTIR GUI on Windows, install additional hardware drivers after activating the environment:

pip install pyvisa pywin32

Quick start

import speclab

# Full emission calibration (reads BB temps from measurement-info.csv)
out = speclab.emcal('/path/to/data/', lab='nau', method='nem')

# Access results
wn     = out['xaxis']          # wavenumber axis (cm⁻¹)
emiss  = out['emiss']          # dict: sample label → emissivity array
labels = out['label']          # ordered list of sample labels

# Spectral mixture analysis
sma_out = speclab.sma(out, endlib='/path/to/library.hdf')

# Merge MIR and FIR emcal runs for the same samples
broadband = speclab.merge(out_mir, out_fir, how='horizontal')

# Transmission calibration from an AutomateFTIR folder
tra_out = speclab.tracal('/path/to/tracal_data/')

# Reflectance calibration
ref_out = speclab.refcal('/path/to/refcal_data/')

# Load and normalise a DV-format spectral library
raw   = speclab.readDVhdf('/path/to/library.hdf')
album = speclab.dv_to_album(raw)

VSWIR reflectance

from speclab import remove_continuum, band_parameters, smooth_spectrum

# wl: wavelength axis in nm, refl: reflectance array (0–1)
refl_sm = smooth_spectrum(wl, refl, method='savgol', window_nm=20, polyorder=3)
cr      = remove_continuum(wl, refl_sm, wl_range=(400, 2500))

# Band depth, centre, FWHM for the 2200 nm Al-OH feature
bp = band_parameters(wl, cr, wl_range=(2150, 2250))
print(bp['depth'], bp['center'], bp['fwhm'])

Instrument presets

list(speclab.INSTRUMENT_PRESETS.keys())
# ['spectrometer', 'nau', 'asu', 'swri', 'tes', 'tes5', 'mtes']

Running the demo

demo.py walks through the full pipeline — emcal → SMA → plots — using the bundled example_data/ and spectral_libraries/ folders.

# Install the package first (only needed once)
pip install -e .

# Run with default settings (single-folder, NEM, rock-forming mineral library)
python demo.py

Key settings at the top of demo.py:

Variable	Default	Description
USE_MULTI_FOLDER	False	True runs emcal on all four example_data/ subfolders and merges
METHOD	'nem'	Emissivity retrieval method: 'nem', 'hullfit', or 'mmd'
ENDLIB_PATH	spectral_libraries/speclib_JFS_rock_forming_minerals.hdf	Endmember library for SMA
USE_SPECLIBVIEWER	False	True opens SpeclibViewer to build a custom library interactively
SHOW_PLOTS	True	Display emcal and SMA result plots
SAVE_RESULTS	False	Write HDF5 + CSV outputs alongside the data

Example data

example_data/ contains:

LWIR emission (Ward's rock standards, NAU)

Subfolder	Rock types
WardRocks_igneous1	Granite, granodiorite, diorite, gabbro, andesite
WardRocks_igneous2	Additional igneous standards
WardRocks_metamorphic	Metamorphic standards
WardRocks_sedimentary	Sedimentary standards

Each subfolder contains sample CSVs, bbhot.CSV, bbwarm.CSV, and a measurement-info.csv notes file with timestamps and instrument channel data.

VSWIR reflectance

Path	Contents
ASD_data/Orochi_analyses_2.csv	ASD field reflectance spectra (Orochi site)
dummy_vswir/	Synthetic VSWIR dataset for testing

Bundled spectral libraries

File	Spectral range	Contents
speclib_JFS_rock_forming_minerals.hdf	LWIR	NAU rock-forming mineral endmember library
ASU_speclib_full_SV.hdf	LWIR	ASU thermal emission spectral library
CSE_wards_rock_library_SV.hdf	LWIR	CSE Ward's rock standards
usgs_splib07_cvASD.hdf	VSWIR	USGS Spectral Library 7 — ASD-convolved variant
crism_speclib.hdf	VSWIR	CRISM spectral library

Spectral library conversion

DV HDF5 → SpectralViewer format

python -m speclab.convert_speclib convert input.hdf output.hdf
python -m speclab.convert_speclib test    input.hdf

USGS splib07 → speclab HDF5

python -m speclab.convert_speclib usgs splib07_dir/ output.hdf
# Restrict to specific chapters and library variant:
python -m speclab.convert_speclib usgs splib07_dir/ output.hdf --source cvASD --chapters Minerals Vegetation

CRISM spectral library → speclab HDF5

python -m speclab.convert_speclib crism input.hdf output.hdf