Urban Flood

Flood mapping from Earth observation (EO) data. This is parked for now but I will look back it again later.

Data will be available from here and ee.ImageCollection(projects/floodmapping-2022/assets/examples) later.

Jobs

Sentinel-1 flood mapping using omnibus test statistics
Urban Sentinel-1 flood mapping as in Chini (2019)
Validation data

SAR Scripts

mortcanty_tutorial.py

Contains functions from this Tutorial to detect changes in Sentinel-1 imagery.

Idai Floodmaps.ipynb

Notebook based on this Tutorial replicating the floodmaps from Cyclone Idai and overlaying them with 2000 mangrove extent maps from Giri (2011). Exports images to Google Earth Engine Project ee.ImageCollection(projects/floodmapping-2022/assets/examples) and Google Drive root directory. Also saves final folium map locally as a .html file.

R Scripts

sen2cor.sh

Just a set of instructions for how to use sen2cor to process Sentinel-2 Level 1C to Level 2A. Not relevant for other machines.

jp2_tiff.sh

Batch translate JPEG2000 files to GeoTIFF so that they can be uploaded to Google Earth Engine.

make_multibands.sh (unfinished)

To generate a single multiband image from a set of .tif files of Sentinel-2 Level 2A imagery. This is not currently working, it generates the images but only creates 3-band images instead of the eight, or so, that it should. I've asked about it on Stack Overflow here so maybe there will be some advice soon.

process_data.R

R code to process data, sample points for training and testing and output into a .csv format.

ml_pipeline.R

R code to train multiple ML models on the datasets and output results.

Useful tutorials

Tutorials to look at when I come back to this:

alisonpeard/urbanflood