Skip to content

Python package that converts between polar and cartesian domain (in images too).

License

Notifications You must be signed in to change notification settings

addisonElliott/polarTransform

Repository files navigation

Build Status Python version PyPi version Documentation Status

Introduction

polarTransform is a Python package for converting images between the polar and Cartesian domain. It contains many features such as specifying the start/stop radius and angle, interpolation order (bicubic, linear, nearest, etc), and much more.

Installing

Prerequisites

  • Python 3
  • Dependencies:
    • numpy
    • scipy
    • scikit-image

Installing polarTransform

polarTransform is currently available on PyPi. The simplest way to install alone is using pip at a command line:

pip install polarTransform

which installs the latest release. To install the latest code from the repository (usually stable, but may have undocumented changes or bugs):

pip install git+https://github.com/addisonElliott/polarTransform.git

For developers, you can clone the polarTransform repository and run the setup.py file. Use the following commands to get a copy from GitHub and install all dependencies:

git clone pip install git+https://github.com/addisonElliott/polarTransform.git
cd polarTransform
pip install .

or, for the last line, instead use:

pip install -e .

to install in 'develop' or 'editable' mode, where changes can be made to the local working code and Python will use the updated polarTransform code.

Test and coverage

Run the following command in the base directory to run the tests:

python -m unittest discover -v polarTransform/tests

Example

Input image:

Cartesian image

import polarTransform
import matplotlib.pyplot as plt
import imageio

verticalLinesImage = imageio.imread('IMAGE_PATH_HERE')

polarImage, ptSettings = polarTransform.convertToPolarImage(verticalLinesImage, initialRadius=30,
                                                            finalRadius=100, initialAngle=2 / 4 * np.pi,
                                                            finalAngle=5 / 4 * np.pi)

cartesianImage = ptSettings.convertToCartesianImage(polarImage)

plt.figure()
plt.imshow(polarImage, origin='lower')

plt.figure()
plt.imshow(cartesianImage, origin='lower')

The result is a polar domain image with a specified initial and final radius and angle:

Polar image

Converting back to the cartesian image results in only a slice of the original image to be shown because the initial and final radius and angle were specified:

Cartesian image

Next Steps

To learn more about polarTransform, see the documentation.

License

polarTransform has an MIT-based license.