Spectral Analysis Utilities

This repository contains scripts and example data for experimenting with NDVI and spectral reflectance models. It is organised into four main directories:

• data/ – input datasets such as spectral CSV files, pigment spectra and a ZIP archive of published leaf litter measurements.
• scripts/ – Python utilities to fit filter curves, simulate NDVI and mix reflectance spectra.
• figures/ – figures produced by the scripts.
• results/ – generated files such as JSON coefficient tables and bundled ZIP archives.

Data

The data directory stores several reference spectra:

• SERAPH_R118_absorbtion (1).json – transmission profile of an optical filter.
• chl_a_122_abs.txt and chl_b_125_abs.txt – chlorophyll absorption data from PhotochemCAD.
• intact_spec.csv, ground_spec.csv, and healthy_leaf.csv – example reflectance measurements.
• 2012-leaf-reflectance-spectra-of-tropical-trees-in-tapajos-national-forest.csv – healthy leaf spectra.
• doi_10_5061_dryad_hdr7sqvrk__v20240426.zip – archive containing additional litter spectra and metadata.
• dataSpec_P5.csv – soil spectra.

Scripts

The scripts folder contains several Python programs:

• bundle.py – downloads pigment spectra, extracts reference data and plots filter transmission alongside chlorophyll reflectance. It creates spectra_bundle.zip and transmission_vs_chl_reflect.png in the results and figures directories.
• r118_filter_fit.py – fits high degree polynomials to the R118 filter transmission curve and saves the coefficients to results/poly14_trans_fit.json. Various plots of the fits and residues are saved to figures/.
• simulate_ndvi.py – computes NDVI values for each row in a reflectance CSV file using the supplied red and NIR filter profiles.
• simulate_filtered_reflectance.py – mixes leaf litter, healthy leaf and ground spectra in different proportions and reports the sensor reading after applying a filter profile.
• simulate_mixed_ndvi.py – similar to the above but outputs NDVI instead of the raw sensor value.

All scripts assume they are executed from the repository root so that relative paths to the data directory work out of the box. Each script prints usage information when run with -h.

Figures and Results

Running the scripts will create PNG figures in figures/ and additional JSON or ZIP artifacts in results/. Several example outputs are already included for reference.

Usage Example

Install the required packages:

pip install numpy matplotlib requests pandas scipy

Then run one of the scripts, for example:

python scripts/simulate_ndvi.py --csv data/intact_spec.csv

Licence

The example data are reproduced from their respective sources. The scripts in this repository are released under the MIT licence.