A project for CSE163, where we compare methods for computing the power of a fractal, and utilize these methods to analyze the world around us.
Install all of the dependencies for this project with
pip3 install -r requirements.txt
geopandas is also a required but I could not find an easy way to install it on windows. If you are on mac or linux you should be able to install this using
pip3 install geopandas
however, on a windows machine, roughness.py will not compile without geopandas I would recomend using a virutal enviroment on edstem, or VirutalBox
All of the data is already included you need to extract each of the foleders composed in images.rar into the main directory, on windows that is usually done with winrar or 7zip. Alternitavely the folders can be downloaded directly from our github page
the directory should look as follows
Fractals\
images\
imgs\
plts\
state_borders\
test_imgs\
compileImages.py
Fractal.py
gz_2010_us_040_00_5m.json
ImageTools.py
ml.py
README.md
requirements.txt
roughness.py
tests.py
Some data was compiled and included by hand, images, state_borders, gz_2010_us_040_00_5m.json, test_imgs. The rest of the data, found in the imgs folder is pre-downloaded, but can also be re-downloaded by running
python3 compileImages.py
this will download images into the imgs folder
A fractal can be analyzed with the following code snippet
from Fractal import Fractal
f = Fractal("path/to/img.png")
f.calculate_power()
There are also the following methods
f.visualize_counting()
f.get_last_power_visualization()
f.set_sample_size()
To run the analysis for the Fractal algorithm on the compiled datasets run
python3 Fractal.py
This will output the images you can see in the report to the plts folder
To analyze images with Machine Learning, run
python3 ml.py
To analyze the borders of US State geospatial data
python3 roughness.py
In order to test all of the modules in this repo run
python3 tests.py
this will error out if there is unexpected behavior in any function