Team Rhoket Scientists for Welthungerhilfe

Video Screencast

This software analyzes a given set of landsat satellite images to show how the land is used. It visualizes the different surface patterns with different colors.

It should help Welthungerhilfe to have strong arguments for disputing with the government of Sierra Leone about shifting land usage patterns.

The project got startet at the 3. Random Hacks of Kindness in Berlin from 2012-06-02 to 2012-06-03 where it got awarded the second prize. In the following months, we re-implemted and extended the application in Python. We closed the voluntary project in September 2013 by demonstrating technical feasibility of an automated analysis of land usage using satellite images to staff from Welthungerhilfe, one of the largest German development cooperation agencies.

• The problem page: http://www.rhok.org/problems/shifting-patterns-land-use-and-availability-land-pujehun-district-southern-sierra-leone

• The solution page: http://www.rhok.org/solutions/vegetation-analyzer

##Run the Application To run the application, simply execute make run.

###Get Help Run python main.py -h to see all avaliable command line options.

##Get Satellite Images

• http://earthexplorer.usgs.gov/ to select images, then grab the bulk downloader application, click an account, open the application and download
• http://landsat.usgs.gov/LandsatLookImages.php
• http://www.uni-kiel.de/forum-erdkunde/unterric/material/einf_fe/* kapitel_4_landsat_typ.htm
• http://www.crwr.utexas.edu/gis/gishydro99/class/gomez/termproj.htm
• http://www.isprs.org/proceedings/XXXV/congress/yf/papers/951.pdf
• http://landsat.usgs.gov/documents/June_2010_Vierling%20-%20Idaho%20Remote%20Sensing%20Activities%20-%20Change%20Assessments.pdf

#For developers

Any help is really appreciated. If you have an idea, great. Just start coding or drop us a line and send us a pull request.

As a developer, you should install the requirements for developers. Just run make req_stable. If you want to regenerate the UI files after editing the a Qt ui file, run make generate_pyqt. This will generate .py files for all ui files.

If you want to have isolated Python environments, we recommend to use virtualenvwrapper.

If you want to freeze the requirements (aka create the requirements files), use pip freeze -r devel-req.txt > stable-req.txt.

Setup on Linux (tested on Ubuntu)

1. sudo apt-get install python-pip for Python dependency management
2. PyQT for the GUI
3. OpenCV for image processing
4. In your .bashrc: export $ PYTHONPATH=/usr/local/lib/python2.6/site-package because Python and opencv only work with Python 2.6 on Ubuntu

##Setup on MacOS X

1. brew install opencv --with-tbb --with-tbb enabes multithreaded image crunching
2. brew install pyqt
3. easy_install pip
4. make req_stable

##Dependencies

• OpenCV for image processing, please compile/install it with WITH_TBB to enable multithreading
• Qt for the GUI
• Python for the code
  • pyqt as a binding for Qt
  • oktest for the unit tests (only for developers)

To install the Python dependencies, simply run make req_stable.