tonywong9527 / colony-counter Goto Github PK

Demo of the script running

Script output images

This repository contains the scrypt, data and outputs used in my project for the COSC428: Comuter Vision paper at the University of Canterbury (21S1).

The University of Canterbury internal paper can be found here for more context.

TL;DR: This project proposes a method to aid in the enumeration of bacterial colonies present on agar plates through use of preprocessing techniques, the Hough Circle Transform and the Watershed Transform.

This script was written using the conda environment and relies on the OpenCV and numpy python libraries, so make sure you have them installed.

Once you have a local copy of the repository, place images of the agar plates in the images/ directory

To run the scrypt, use the following command

./counter.py <input-file> <method-to-use>

The supported methods are:

• h: Hough Circle Transform.
• w: Watershed Transform.

The input file is assumed to be in the images/ directory, so you don't have to include the directory in the filename.

For example, when using "plate1.jpg" you don't need to run the scrypt with images/plate1.jpg, just plate1.jpg.

Like so:

./counter.py plate1.jpg h

If you'd prefer not to use the console, you can use the following call to main:

if __name__ == '__main__':

    # Uncomment this line if not running through terminal,
    # only change the second and third elements
    main(['counter.py', 'plate1.jpg', 'h'])

    #main(sys.argv)

The script will run the same using this method, whatever is easier.

Kayle Ransby - [email protected]

Source Code Link: https://github.com/krransby/colony-counter

Report Link: Coming soon

• University of Canterbury
• Richard Green (Supervisor)
• images/HQ
• plate1.jpg
• plate2.jpg