Skip to content

Developed deep learning based architecture model which perform semantic segmentation for post-flood images using VGG16 pre-trained model and Unet architecture.

Notifications You must be signed in to change notification settings

sidd1915/Semantic-Segmentation-for-Flood-Images

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SEMANTIC SEGMENTATION OF DRONE UAV IMAGES TO IMPROVE POST-DISASTER MANAGEMENT

Semantic segmentation of drone UAV images to improve disaster management by detecting and segmenting flood-affected areas. The model is trained on the Floodnet dataset.

Table of Contents

Project Description

This project aims to enhance disaster management by using a deep learning model to perform semantic segmentation on drone images. The model helps in identifying and segmenting flood-affected areas, providing valuable information for response and recovery efforts. It can classify the given image into 10 classes - ('Background':0, 'Building-flooded':1, 'Building-non-flooded':2, 'Road-flooded':3, 'Road-non-flooded':4, 'Water':5, 'Tree':6, 'Vehicle':7, 'Pool':8, 'Grass':9)

Demonstration

The Streamlit app provides an interface to upload images and view the segmentation results in real-time. Streamlit Demonstration - Click here

Dataset

The model is trained on the Floodnet dataset. You can download the dataset from Floodnet Dataset. Please also refer to the research paper paper .

Results

We used Iou per each class to measure the accuracy of our model. image

License

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

Acknowledgements

We would like to thank the developers of the Floodnet dataset and the contributors to open-source deep learning libraries that made this project possible.

Citations for papers referred - @ARTICLE{9460988, author={Rahnemoonfar, Maryam and Chowdhury, Tashnim and Sarkar, Argho and Varshney, Debvrat and Yari, Masoud and Murphy, Robin Roberson}, journal={IEEE Access}, title={FloodNet: A High Resolution Aerial Imagery Dataset for Post Flood Scene Understanding}, year={2021}, volume={9}, number={}, pages={89644-89654}, doi={10.1109/ACCESS.2021.3090981} }

@article{rahnemoonfar2020floodnet, title={FloodNet: A High Resolution Aerial Imagery Dataset for Post Flood Scene Understanding}, author={Rahnemoonfar, Maryam and Chowdhury, Tashnim and Sarkar, Argho and Varshney, Debvrat and Yari, Masoud and Murphy, Robin}, journal={arXiv preprint arXiv:2012.02951}, year={2020} }

About

Developed deep learning based architecture model which perform semantic segmentation for post-flood images using VGG16 pre-trained model and Unet architecture.

Topics

Resources

Stars

Watchers

Forks