What is ContrAI?
The ContrAI repository provides an easy-to-use framework for performing contrail detection using the neural network models evaluated in our research paper, enabling users to seamlessly access satellite data, run inference on new satellite imagery, and export georeferenced outputs for visualization, cross-validation with other data sources, or comparison with alternative modeling approaches; the repository is under active development and will continue to expand with additional features, tools, and datasets to support advanced contrail analysis and satellite image processing workflows.
ABOUT
- License: This project is licensed under the Apache 2.0 License.
- Reference:
- I. Ortiz, J. García-Heras, A. Jafarimoghaddam and M. Soler, “Robust Evaluation of Neural Networks Trained on the OpenContrails Dataset,” in IEEE Transactions on Geoscience and Remote Sensing, vol. 63, pp. 1-17, 2025, Art no. 4112717, doi: 10.1109/TGRS.2025.3629628.
- Support: Support on all technical questions will be provided by Irene Ortiz (irortiza@ing.uc3m.es)
- Core developer team: Irene Ortiz, Javier García-Heras and Manuel Soler