Skip to main content

Adjeff

Adjeff simulates and corrects the adjacency effect in Sentinel-2 imagery: the light that the atmosphere scatters from neighbouring pixels into the one observed.

  • Forward simulation: top-of-atmosphere reflectance of any surface, by GPU Monte Carlo radiative transfer (Smart-G).
  • PSF learning: the atmospheric point spread function $P$, fitted per band and atmospheric state as an analytical kernel.
  • Correction: surface reflectance recovered from an observed scene, with the learned PSF.

$$\rho_{toa} = \rho_{atm} + T^\downarrow \frac{T^\uparrow_{dir},\rho_s + T^\uparrow_{dif},(\rho_s \ast P)}{1 - s,(\rho_s \ast P)}$$

Installation

Adjeff is managed with pixi: Smart-G and part of its dependencies come from conda-forge.

git clone https://github.com/walcark/Adjeff.git && cd Adjeff
pixi install -e gpu          # or -e cpu, without Smart-G simulations
export SMARTG_DIR_AUXDATA=/path/to/smartg/auxdata

Simulations need a CUDA 12.6 GPU and the Smart-G auxiliary data. Analysis, PSF models and the 5S formulas run on CPU.

Example

import adjeff
from adjeff.core import KingPSF, S2Band, disk_image_dict
from adjeff.modules.models import Unif2Surface
from adjeff.optim import TrainingImages

bands = [S2Band.B03]
cfg = adjeff.make_full_config(bands, aot=0.2, rh=50.0, h=0.0, sza=30.0, vza=5.0)

# Bright disks of 1, 5 and 50 km on a dark ground, 120 m pixels.
disks = [
    disk_image_dict(radius=r, res_km=0.12, rho_max=0.5, bands=bands, n=1999)
    for r in (1.0, 5.0, 50.0)
]

# Simulate their radiative terms, rho_toa and rho_unif (GPU).
disks = adjeff.run_forward_pipeline(disks, **cfg)

# Fit the King PSF that recovers rho_s from rho_unif.
model = adjeff.make_model(
    Unif2Surface, KingPSF, bands, res_km=0.12, n=1999,
    init_parameters={"sigma": 0.5, "gamma": 2.0},
)
tree = adjeff.fit(model, TrainingImages(disks))

# Correct a scene carrying rho_unif and the radiative terms.
corrected = adjeff.apply_psf(disks[0], tree, S2Band.B03)

adjeff.fit_psf(scene, bands, KingPSF, ...) does the same in one call, at the mean atmosphere of a loaded product.

Documentation

User guide Concepts, configuration, samplers, PSF, fitting, cache, logging
Notebooks Six worked examples, from a synthetic scene to a PSF fitted and measured
Changelog What changed, and why
Roadmap Open questions and planned work

Development

pixi install -e dev
pixi run -e dev all                    # format, lint, mypy --strict, tests
pixi run -e dev-gpu test-integration   # Smart-G tests, on GPU

License

Apache-2.0.

Metadata

Release files for adjeff 1.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for adjeff 1.2.0
File Size Uploaded Provenance Actions
adjeff-1.2.0.tar.gz 1.6 MB With provenance Details for adjeff-1.2.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for adjeff 1.2.0
File Interpreter ABI Platform Actions
adjeff-1.2.0-py3-none-any.whl Python 3 none any Details for adjeff-1.2.0-py3-none-any.whl

Total release size: 1.7 MB

Release files / adjeff-1.2.0.tar.gz

Download URL adjeff-1.2.0.tar.gz
Size 1.6 MB
Tags Source
SHA-256 checksum
How to use checksums
2496810eec026b97b8cc6301a4ebcc7245b0a39c295e7197a4a49a559a84ea9b
BLAKE2b-256 checksum
How to use checksums
04f8c7dc9d17efcf388ea43b7aa346f5d733f364d51762a819138b38b733635c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 13, 2026.

Transparency log

Release files / adjeff-1.2.0-py3-none-any.whl

Download URL adjeff-1.2.0-py3-none-any.whl
Size 138.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
7fb891ff22c5ab6eca560504e2d229a231e342fd5ea106d3d1cfc9941b7f0787
BLAKE2b-256 checksum
How to use checksums
a6fe3306055ee94f02ab363844330a916750318ee184acdc32730ed7a215bd12
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 13, 2026.

Transparency log

Security for adjeff 1.2.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for adjeff-1.2.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 13, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page