This service is intended to be used on a desktop computer for ease

Welcome to the SatCropper UNHCR Service

In collaboration with the University of Essex computer vision team – via its Human Rights Big Data and Technology Initiative (HRBDT). Alongide with the inputs of Prof. Klaus McDonald-Meier, Prof. Geoff Gilbert, and Researchers/Developers: Bruno Queiros Arcanjo, Griogorios Kalliatakis and Ciro Iván Agulló - we have built an alpha version for remote sensing (admin level 2) using Python and Landsat 8 products for researchers and other humanitarians to find the nexus between climate and forced displacement. Still in pilot version, this application allows the user to run analysis on satellite imagery to extract information about extreme weather events (e.g., droughts/floods) and derive a calculation for anticipatory action (e.g. building predictive models with that data).

This project will run on your computer, download large satellite images, crop these images, perform calculations and extract the data for you to analyze.

Setup and KML lookup:

  1. Download and install docker: Follow the instructions here
  2. Use the map on the right to download the KML file of the region you're going to perform the analysis:
    1. All african countries are highlighted on the map
    2. Click on a country to highlight admin2 levels
    3. At any point, you can click the R button on the top left corner to reset the country view
    4. Click again on any admin2 level to start the download process
    5. A KML file representing the admin2 level boundaries will be downloaded
  3. Locate the recently downloaded KML file (usually in downloads folder)
  4. We recommend creating an empty folder somewhere in the system
  5. Place this file there and remember this location
  6. Open a terminal or shell within that location

Register on the USGS Service:

  1. You can register an account here: follow the instructions on this link
  2. After your credentials have been successfuly created you can proceed to the next step

Commands and explanation:

Some context: We're going to execute a container that has all dependencies and python scripts to perform all the hardwork, however we still need to pass some context to this container, that context being some parameters, the KML file itself and temporary files being generated during the process. Some bash or shell proefficiency is nice to have but we're going to try to explain all steps so you don't have to understand the details.

Please ensure you're placed on the same directory that contains the KML file. Use ls or dir commands to always check for these files.

Format: docker run -v `pwd`:/shared public.ecr.aws/r6y8h7j3/satcropper-public <arguments> is a pattern that will be repeated trough all the process, we're only going to change the arguments being sended to the container on each step.

There are different ways to pinpoint to volumes on the 'working directory' depending your OS, for instance in PowerShell, you use ${PWD}, which gives you the current directory. So the command will be docker run -v ${PWD}:/shared public.ecr.aws/r6y8h7j3/satcropper-public <arguments> instead.

  1. $ docker run -v `pwd`:/shared public.ecr.aws/r6y8h7j3/satcropper-public search_scenes $your_user $your_password $kml_filename $month $results
  2.           $your_user: represents the username you've registered on the USGS services
              $your_password: being the password you've registered on the USGS services
              $kml_filename: has to be the filename without the .kml extension
              $month: has to be a number representing the month you want to look up
              $results: usually 20 represents a year, so increasing the number may include previous years
            
  3. $ docker run -v `pwd`:/shared public.ecr.aws/r6y8h7j3/satcropper-public download_scenes $your_user $your_password $kml_filename
  4.           $your_user: represents the username you've registered on the USGS services
              $your_password: being the password you've registered on the USGS services
              $kml_filename: has to be the filename without the .kml extension
            
  5. $ docker run -v `pwd`:/shared public.ecr.aws/r6y8h7j3/satcropper-public extract_scenes $kml_filename
  6.           $kml_filename: has to be the filename without the .kml extension
            
  7. $ docker run -v `pwd`:/shared public.ecr.aws/r6y8h7j3/satcropper-public process_scenes $kml_filename
  8.           $kml_filename: has to be the filename without the .kml extension