Open-source codes and datasets from the Olugboji Planetary Imaging and Earth Hazards Lab, organized by the three stages of the lab’s methods program. Each tool links to its code, its archived release and the paper to cite.

Stage 1 · Advanced Signal Denoising
CRISP-RFFADERRFDereverbSHARP-SS
Stage 2 · AI-Powered Seismic Refinery
AkiNetiRADnet
Stage 3 · Interpretable AI
Spec2VEC
Two spectrograms feeding the Spec2VEC engine (Figure 1 of the paper)
Featured · Stage 3 · Interpretable AI
Spec2VEC
Interpretable SPECtral representations of planetary signals in VECtor spaces

Turns spectrograms into interpretable feature vectors that sort marsquakes, eruptions and other planetary signals with few labels.

Jupyter Notebook · MIT license · updated Aug 2026
Dockerdocker pull urseismogate.earth.rochester.edu/spec2vec:latest

Stage 1 · Advanced Signal Denoising

Recovering reliable structure from the sparsest, noisiest seismic records.

CRISP-RFClean Receiver-function Imaging with SParse Radon Filters

Separates direct mantle conversions from crustal multiples and noise in receiver functions using a sparse Radon transform.

MATLAB · updated Jan 2024
FADERFast and Automated Detection and Elimination of Echoes and Reverberations

Detects and removes reverberations from sediments, oceans and glaciers in any seismic recording, without prior structure.

MATLAB · MIT license · updated Jan 2024
RFDereverbSediment reverberation filter for ocean-bottom receiver functions

Removes sediment reverberations from ocean-bottom receiver functions; the forerunner of FADER.

MATLAB · GPL-3.0 license · updated Sep 2021
SHARP-SSSparse High-resolution Algorithm for Reflection Profiling with SS-waves

Probabilistic deconvolution that separates SS waves from faint precursors to image mantle discontinuities with sparse data.

MATLAB · updated Nov 2024

Stage 2 · AI-Powered Seismic Refinery

Physics-informed AI that turns raw vibration into high-fidelity structural information.

AkiNetPhysics-informed neural network for noise-based surface-wave dispersion

Physics-informed neural network that extracts surface-wave dispersion from ambient-noise correlations.

Jupyter Notebook · updated Sep 2025
iRADnetDeep algorithm unrolling for the inverse Radon transform

Deep algorithm unrolling that separates reflected wavefields for seismic migration.

MATLAB · GPL-3.0 license · updated Aug 2025

Datasets & models

Open models and datasets behind the lab's continental and mantle maps.

ADAMAAfrica's crustal architecture: shear-velocity model, crustal taxonomy and phase-velocity maps

Africa's crustal shear-velocity model, crustal taxonomy and ambient-noise phase-velocity maps, built from the largest catalog of short-period surface-wave measurements.

MATLAB · updated Apr 2024
ADAMA.SWaveShort-period surface-wave dispersion dataset for Africa

About 100,000 Love- and Rayleigh-wave phase-dispersion measurements from ambient noise across Africa, for testing and updating crustal models.

MATLAB · updated Jan 2024
US Mantle TaxonomyTaxonomy of upper-mantle stratification beneath the US

Dataset and codes for classifying upper-mantle layering beneath the United States from receiver functions.

Jupyter Notebook · updated Mar 2024

All datasets and models, with downloads and citation details: Open Data

Earlier tools

RFMigHarmonicReceiver-function harmonic stacking and migration (C++)

Command-line receiver-function utility: harmonic stacking, depth migration and multi-layer H-K stacking.

C++ · updated Jan 2024

All lab code on GitHub

github.com/URseismology 26 public repositories: tools, datasets, notebooks and course materialBrowse →

When you use a tool, please cite its paper and its archived release (Release DOI). Questions and bug reports are welcome as issues on each tool’s GitHub page.

Leave a Reply

Your email address will not be published. Required fields are marked *


The reCAPTCHA verification period has expired. Please reload the page.

This site uses Akismet to reduce spam. Learn how your comment data is processed.