/hough-nn

Parallel Hough Transform & Nearest Neighbours image algorithms

Primary LanguagePythonMIT LicenseMIT

hough-nn

Parallel Hough Transform & Nearest Neighbours image algorithms

Steps

  • Save desired images to process under images/original: they should be ideally simple images (such as vectorized ones) where objects are clearly delimited, and shading doesn't have much impact.

    • Additionally, the background must be white - the program will consider white pixels as not being part of any objects
  • Run image preprocessing: preprocess_images.py

    • Will preprocess all images under the aforementioned folder
  • To run hough transform: mpiexec -n 2 python hough.py

    • Store results as a csv file in results/hough
  • plot results using plot_from_csv.py, specifying desired path in script