CryoTrack: Iceberg Tracking and Lineage Analysis

This repository holds the code described in the paper: Icebergs, Jigsaw Puzzles and Genealogy: Automated multi-generational iceberg tracking and lineage reconstruction Submitted to The Cryosphere

It contains scripts and tools for tracking icebergs using satellite imagery, constructing tracklets, establishing generational linkages, and evaluating tracking performance. The workflow is designed for use with the CI2D3 database (https://wirl.carleton.ca/research/ice/ice-islands/ci2d3/ .

Repository Structure

• preprocessing/ — Scripts for preparing shapefiles and converting to btrack-compatible JSON.
• tracklet_generation/ — Scripts for generating tracklets using btrack.
• generational_linking/ — Scripts for linking tracklets across generations and constructing iceberg lineages.
• evaluation/ — Scripts for evaluating and visualising tracking results, including traccuracy integration.
• data/ — Input shapefiles, intermediate files, and results (not included in repo).
• notebooks/ — Example Jupyter notebooks demonstrating the full workflow.

Main Workflow

1. Preprocessing:
   Convert raw shapefiles to JSON for btrack input.
2. Tracklet Generation:
   Run btrack to generate initial tracklets.
3. Vector Field Construction:
   Build interpolated velocity fields for spatial priors.
4. Generational Linking:
   Link tracklets across generations to reconstruct iceberg lineages.
5. Evaluation:
   Assess tracking performance using built-in metrics and traccuracy.

Requirements

• Python 3.8+
• btrack
• traccuracy

See environment.yml or use mamba list to view all dependencies.

Usage

For example workflow see [tracking.ipynb] ()

License

This project is licensed under the MIT License.
See the LICENSE file for details.

Acknowledgements

• The CI2D3 database contributors and the WIRL group at Carleton University for providing the iceberg data.
• btrack for the tracking engine.
• traccuracy for parts of the evaluation framework pertaining to the Cell Tracking Challenge
• [inshade] (https://github.com/HBKUVisCommunity/inshade/tree/master)
• The open-source Python community, especially the developers of geopandas, pandas, numpy, matplotlib, shapely
• Funding and support from NERC and collaborating institutions.