hyperspec

Fast hyperspectral processing for remote sensing and machine learning. Rust core with Python bindings.

Built by Hyperstate — foundation models for Earth intelligence.

Why hyperspec?

Hyperspectral imagery produces hundreds of bands per pixel, orders of magnitude more than RGB or multispectral. Most existing tools are slow, have heavy system dependencies, and were not designed for modern ML training pipelines.

hyperspec provides:

• Fast Rust implementations of standard hyperspectral algorithms
• A SpectralCube abstraction that keeps data, wavelengths, and metadata together
• Zarr V3 storage with chunking optimized for ML dataloaders
• Python bindings with efficient NumPy interop

Install

Python:

pip install hyperspec-py
# or
uv add hyperspec-py

Rust:

cargo add hyperspec

or

[dependencies]
hyperspec = "0.8"

Quickstart

from hyperspec import (
    read_envi, write_zarr, read_zarr_window,
    band_stats, covariance,
    mnf_denoise, savitzky_golay, derivative, continuum_removal, normalize_zscore,
    sam, ndvi, pca,
)

# --- I/O ---

cube = read_envi("scene.hdr")

# --- Exploration ---

stats = band_stats(cube)        # .mean, .std, .min, .max, .valid_count
cov = covariance(cube)          # (bands, bands) covariance matrix

# --- Preprocessing ---

clean = mnf_denoise(cube, n_components=20)  # denoise first
smooth = savitzky_golay(clean, window=7, polyorder=2)
d1 = derivative(smooth, order=1)
cr = continuum_removal(clean)
normed = normalize_zscore(clean)

# --- Analysis ---

reference = normed.spectrum(50, 50)
angles = sam(normed, reference)             # spectral similarity
veg = ndvi(normed, nir_band=90, red_band=55)  # vegetation index
pca_result = pca(normed, n_components=10)   # dimensionality reduction

# --- Storage ---

write_zarr(normed, "chips.zarr", chunk_shape=(normed.bands, 256, 256))
tile = read_zarr_window("chips.zarr", bands=(0, normed.bands), rows=(0, 256), cols=(0, 256))

SpectralCube

Category	Methods
Dimensions	bands, height, width, shape
Data access	data(), wavelengths(), fwhm(), nodata
Pixel access	spectrum(row, col)
Band access	band(index), band_at(nm), band_nearest(nm)
Wavelength lookup	wavelength(index), wavelength_index(nm), nearest_band_index(nm)
Statistics	mean_spectrum()
Subsetting	sel_bands([i, j]), sel_wavelengths(min, max)

I/O

Read

Format	Function
ENVI	read_envi(path)
Zarr V3	read_zarr(path)
Zarr V3	read_zarr_with_options(path, ...)
Zarr V3	read_zarr_window(path, bands, rows, cols)
Zarr V3	zarr_cube_shape(path)

Write

Format	Function
ENVI	write_envi(cube, path)
Zarr V3	write_zarr(cube, path)

Algorithms

Exploration

Operation	Function
Per-band stats	band_stats(cube) → BandStats
Covariance matrix	covariance(cube)
Correlation matrix	correlation(cube)

Preprocessing

Operation	Function
MNF denoise	mnf_denoise(cube, n_components)
MNF	mnf(cube, n_components)
Savitzky-Golay smoothing	savitzky_golay(cube, window, polyorder)
Derivative spectra	derivative(cube, order)
Continuum removal	continuum_removal(cube)
Min-max normalization	normalize_minmax(cube)
Z-score normalization	normalize_zscore(cube)
Resample	resample(cube, target_wl, method)

Analysis

Operation	Function
Spectral Angle Mapper	sam(cube, reference)
Normalized difference	normalized_difference(cube, a, b)
Band ratio	band_ratio(cube, a, b)
NDVI	ndvi(cube, nir_band, red_band)
PCA	pca(cube, n_components)

Architecture

crates/hyperspec/           # Pure Rust library → crates.io
└── src/
    ├── cube.rs             # SpectralCube type
    ├── error.rs            # HyperspecError
    ├── io/
    │   ├── envi.rs         # ENVI read/write (BSQ, BIL, BIP)
    │   └── zarr.rs         # Zarr V3 read/write/window
    ├── linalg.rs           # Matrix ops (faer): eigen, covariance, tiled GEMM
    └── algorithms/
        ├── sam.rs
        ├── continuum.rs
        ├── indices.rs
        ├── pca.rs
        ├── mnf.rs
        ├── resample.rs
        ├── stats.rs
        ├── normalize.rs
        ├── derivative.rs
        └── smooth.rs

pyo3-hyperspec/             # PyO3 bindings → PyPI: hyperspec-py
├── src/lib.rs
└── python/hyperspec/
    └── __init__.py

• Rust 2024 edition, using ndarray, rayon, faer, zarrs, thiserror
• Minimal system dependencies — no LAPACK or GDAL
• Python 3.12+, NumPy bindings via PyO3

Roadmap

• Cloud-native Zarr stores (Azure, GCS, S3)
• More algorithms and utilities

Development

uv venv && source .venv/bin/activate
uv pip install maturin pytest
cargo test
maturin develop
pytest

License

Apache-2.0