Fast (ish) Checkboard Demodulation

Fast Checkerboard Demodulation (FCD) for Synthetic Schlieren imaging, in python

Almost completely based on the work from Sander Wildeman: https://github.com/swildeman/fcd

fcd_diagram_image

Usage

To apply the FCD to a single or series of images, run fcd.py as a program:

Run python fcd.py --help to see how to use it.

usage: fcd.py [-h] [--output-format {tiff,bmp,png,jpg,jpeg}] [--skip-existing]
              output_folder reference_image definition_image
              [definition_image ...]

positional arguments:
  output_folder
  reference_image
  definition_image      May contain wildcards

optional arguments:
  -h, --help            show this help message and exit
  --output-format {tiff,bmp,png,jpg,jpeg}
                        The output format (default: tiff)
  --skip-existing       Skip processing an image if the output file already
                        exists (default: False)

Example on generated test image

If you don't have images to apply FCD on, you can generate one with generate_test_images.py: run python generate_test_images.py ripples

This program will generate a reference image, a distorted image and a height field to test against. If you like smileys, you can replace ripples by smiley.

If you like to change some parameters (e.g. pattern frequency), run python generate_test_images.py --help to see what can be changed and how.

To apply the FCD, run: python fcd.py test_output "test image reference.tiff" "test image displaced.tiff"

This will apply the FCD with the output file test_output/test image displaced.tiff, which should resemble the file test image height field.tiff

Dependencies

Python >= 3.7 and the following PIP packages:

numpy
scipy
imageio
more_itertools
scikit-image

TODO

Improve the performance. The current implementation is not very fast

About FCD

Information on the FCD technique and the employed algorithms can be found in: Wildeman S., Real-time quantitative Schlieren imaging by fast Fourier demodulation of a checkered backdrop, Exp. Fluids (2018) 59: 97, https://doi.org/10.1007/s00348-018-2553-9, or https://arxiv.org/abs/1712.05679 (preprint)

kaspervn / pyfcd

readme