slum-orei

Semantic segmentation of slum areas from high-resolution RGB images using three different architectures: U-Net, FPN, and Linknet, implemented in the segmentation_models. Slum prediction implementation using Smoothly-Blend-Image-Patches.

This repo provides RGB image and annotation sample for Jakarta. Users need to modify the directory path accordingly.

Requirements

• segmentation_models
• keras == 2.10 or tensorflow == 2.10
• numpy
• rasterio
• matplotlib
• sklearn
• patchify
• albumentations
• Smoothly-Blend-Image-Patches

Segmentation Models installation

PyPI stable package

pip install -U segmentation-models

PyPI latest package

pip install -U --pre segmentation-models

Source latest version

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

Semantic segmentation of slum areas

Image patching

python image_patching.py

Split train/val data

python split_train_val.py

Train/val

python train_val_slum.py

Slum prediction

python predict_slum.py

Citing

@article{Lumban-Gaol2023,
    author = {Lumban-Gaol, Y. A. and Rizaldy, A. and Murtiyoso, A.},
    title = {COMPARISON OF DEEP LEARNING ARCHITECTURES FOR THE SEMANTIC SEGMENTATION OF SLUM AREAS FROM SATELLITE IMAGES},
    journal = {The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences},
    volume = {XLVIII-1/W2-2023},
    year = {2023},
    pages = {1439--1444},
    url = {https://isprs-archives.copernicus.org/articles/XLVIII-1-W2-2023/1439/2023/},
    doi = {10.5194/isprs-archives-XLVIII-1-W2-2023-1439-2023}
    }