__              ____              
  / /_____  ____  / / /_  ____  _  __
 / __/ __ \/ __ \/ / __ \/ __ \| |/_/
/ /_/ /_/ / /_/ / / /_/ / /_/ />  <  

your personal image processing toolset!

Installing / Building

This toolset is heterogeneous; some are written in C with bindings to libraries like opencv and tesseract, others are in the python scripting language and rely on python packages like numpy, python-opencv, pillow, and scipy.


For the python tools, it may be useful to create a "virtualenv". Otherwise you might eventually have trouble with version conflicts with other tools. Option A describes this approach. Alternatively, you can install the packages in your system (option B).

A. Installing Python dependencies in a venv

  1. Create one:

    virtualenv venv

  2. Activate it:

    source venv/bin/activate

  3. Use pip to install (nb you don't need sudo since you "own" the venv).

    pip install scipy pillow numpy

B. Installing Python dependencies to the system

sudo pip install scipy pillow numpy


Some of the C-based tools like contours and channel have a simple makefile and build simply with:

cd contours

Other tools like texture and houghlines use cmake. Basically, if there is a file called "CMakeLists.txt" the build process uses cmake. An example of building texture would be:

cd texture
mkdir build
cd build
cmake ..


Image Gradient

python python/imagegradient.py foo.png --output foo.gradient.png --svg foo.gradient.svg

Converts foo.png (input) into two outputs.


Nov 24 2016

Incorporate orderings functions: * Way to talk of "items" and collect metadata in a single JSON. * Scripts to generate (leaflet) maps based on metadata. * Cross linkages with other visualisations (such as the map)

What's more complex though is that now each layer is based on a resolution which may (or may not) change the resulting order. So the orderings are now PER ZOOM LEVEL (which is interesting).

ALSO need to make a scraper for RMAH -- do it promiscuously?!