Rapid mapping of flood inundation by deep learning-based image super-resolution

# Rapid mapping of flood inundation by deep learning-based image super-resolution

# Developer: Wenke Song

# The University of Hong Kong

# Contact email: [e-mail]

# MIT License

# Copyright (c) 2024 songwk0924

There are two folders in the compressed file: Trained_model and Test_cases:

(1) Trained_model

      model_d_DenseUnet.pth, for predicting the maximum water depth;

      model_v_DenseUnet.pth, for predicting the maximum velocity.

(2) Test_cases

      Test_d_r1.npy, Test_d_r2.npy, Test_d_r3.npy: Input features for predicting maximum water depth of rainfall events r1-r3;

      Test_v_r1.npy, Test_v_r2.npy, Test_v_r3.npy: Input features for predicting maximum velocity of rainfall events r1-r3;

      bathy_mat_5m_0p.csv: Elevation data to create mask layer;

      Fine_grid_flood_maps (2DSWEs):

             hmax_r1.asc, hmax_r2.asc, hmax_r3.asc, maximum water depth simulated by 2DSWEs of rainfall events r1-r3;

             velmax_r1.asc, velmax_r2.asc, velmax_r3.asc, maximum velocity simulated by 2DSWEs of rainfall events r1-r3;

The aforementioned data will be used as input for model prediction (Prediction.py). 

https://github.com/songwk0924/Flood-inundation-mapping