Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/sidgan/etci-2021-competition-on-flood-detection

Experiments on Flood Segmentation on Sentinel-1 SAR Imagery with Cyclical Pseudo Labeling and Noisy Student Training
https://github.com/sidgan/etci-2021-competition-on-flood-detection

arxiv-papers deep-learning floods noisy-student pseudo-labeling python pytorch segmentation semi-supervised-learning

Last synced: 4 days ago
JSON representation

Experiments on Flood Segmentation on Sentinel-1 SAR Imagery with Cyclical Pseudo Labeling and Noisy Student Training

Awesome Lists containing this project

README 

        
# Flood Detection Challenge



This repository contains code for our submission to the **ETCI 2021 Competition on Flood Detection** ([Leaderboard](https://competitions.codalab.org/competitions/30440), [Homepage](https://nasa-impact.github.io/etci2021/)) (Winning

Solution #2). 



Accompanying paper: [Flood Segmentation on Sentinel-1 SAR Imagery with Semi-Supervised Learning](http://arxiv.org/abs/2107.08369).



by Sayak Paul\*, Siddha Ganju\*.



(\*) equal contribution.





	





**Update (May 17, 2022)**: Orange Business Services, France optimized our public code and made it 2.5 times faster for inference, now inference takes 1.22 seconds over a 63,000 square kilometers tile! Together with Balcony Labs, Oakland, California, an integrated solution was developed which can deliver notifications to mobile devices for early warning and disaster management and has been operationally deployed in Afghanistan, Bangladesh, Mexico, Haiti, and Ukraine. [Read here](https://siliconvalley.orange.com/en/news/rapid-disaster-detection-and-response-a-climate-action-collaboration/)



**Update (May 1, 2022)**:  NVIDIA Deep Learning Institute leveraged this code and produced a new, free, educational course called Disaster Risk Monitoring Using Satellite Imagery, where one can learn to build and deploy a deep learning model to automate the detection of floods using satellite imagery. Check it out [here](https://courses.nvidia.com/courses/course-v1:DLI+S-ES-01+V1/?ncid=so-nvsh-624169#cid=dli03_so-nvsh_en-us)



**Update (April 23, 2022)**: We got a [shoutout from PyTorch](https://twitter.com/PyTorch/status/1517662119700795392).



**Update (December 11, 2021)**: We are releasing a [presentation deck](https://docs.google.com/presentation/d/1CMNK5FuNht_m6owi8lSHd0ksa50JVPeYU88XF39AzA4/edit?usp=sharing), a [presentation video](https://recorder-v3.slideslive.com/?share=52682&s=0258146f-46b5-4696-83a8-60ab25be06c7), and [a poster](https://docs.google.com/presentation/d/1zza2g-9YTsifrq3v10YWz5tbNcsAV0ILANTfuz-IgSI/edit?usp=sharing) of our work. 



**Update (October 26, 2021)**: Our work made it to the following NeurIPS 2021 workshops: [_AI for Science: Mind the Gaps_](https://ai4sciencecommunity.github.io/), [_Tackling Climate Change with Machine Learning_](https://www.climatechange.ai/events/neurips2021), 

[_Women in ML_](https://wimlworkshop.org/sh_events/wiml-neurips2021/), and [_Machine Learning and the Physical Sciences_](https://ml4physicalsciences.github.io/).



## Team 



* [Siddha Ganju](http://sidgan.github.io/siddhaganju)

* [Sayak Paul](https://sayak.dev)



## Executing the code



We executed the scripts and notebooks on a [Vertex AI](https://cloud.google.com/vertex-ai) Notebook instance. The instance has four

Tesla V100 GPUs and its base configuration is [`n1-standard-16`](https://cloud.google.com/compute/docs/machine-types).



We use Python 3.8 and PyTorch 1.9. Apart from the requirements specified in `requirements.txt` you'd need to install the following

as well to run the scripts and notebooks:



```shell

$ pip install git+https://github.com/qubvel/segmentation_models.pytorch

$ pip install git+https://github.com/lucasb-eyer/pydensecrf.git

```



Our scripts and notebooks make use of mixed-precision training (via [`torch.cuda.amp`](https://pytorch.org/docs/stable/notes/amp_examples.html)) and distributed training (via [`torch.nn.parallelDistributedDataParallel`](https://pytorch.org/docs/stable/generated/torch.nn.parallel.DistributedDataParallel.html)). With this combination we are able to achieve significant boosts in the overall model training time.



**Execution instructions for the scripts (`src`) and notebooks (`notebooks`) are provided in their respective directories.** Download the data from the [website](https://competitions.codalab.org/competitions/30440) after agreeing to their license agreements.



## Pre-trained weights



For complete reproducibility, we provide the pre-trained weights [here](https://github.com/sidgan/ETCI-2021-Competition-on-Flood-Detection/releases/download/v1.0.0/pretrained_weights.tar.gz). With these weights and the workflow depicted in the notebooks and scripts one should be able to get an IoU of ~0.76 (as per the [competition leaderboard](https://competitions.codalab.org/competitions/30440#results)) on the test set provided at the competition.



## Results



You can verify the reported results [here](https://competitions.codalab.org/competitions/30440#results). Just switch to "Test (Phase 2)" after

clicking the link.





	




## FAQ



- [Create a new Notebooks instance](https://cloud.google.com/notebooks/docs/create-new)



## Acknowledgements



* We are grateful to the [ML-GDE program](https://developers.google.com/programs/experts/) for providing GCP credits to support our experiments. 

* Thanks to [Charmi Chokshi](https://in.linkedin.com/in/charmichokshi), and domain experts Shubhankar Gahlot, May Casterline, Ron Hagensieker, Lucas Kruitwagen, Aranildo Rodrigues, Bertrand Le Saux, Sam Budd, Nick Leach, and, Veda Sunkara for insightful discussions.



## Citation



```

@inproceedings{paul2021flood,

    title   = {Flood Segmentation on Sentinel-1 SAR Imagery with Semi-Supervised Learning},

    author  = {Sayak Paul and Siddha Ganju},

    year    = {2021},

    URL = {https://arxiv.org/abs/2107.08369},

    booktitle = {NeurIPS Tackling Climate Change with Machine Learning Workshop}

}

```