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

I/O

Read

Write

Algorithms

Exploration

Preprocessing

Analysis

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