Optical Micromouse Maze Recognition

This project aims to ease the process of creating maze files simply using images of micromouse mazes.

Usage

Install it with:

pip install ommr

Run it with:

ommr input_image.png

The generated output file complies with the standard micromouse maze text format. For more information visit micromouseonline/mazefiles.

Image requirements

This software is currently very limited, so the input image must meet some requirements:

• There should be a good contrast between background and walls.
• The maze image must be regular. That means, maze boundaries should look all straight and parallel or perpendicular to each other.
• The maze image must be cropped manually so that top and left boundary walls are included and bottom and right boundary walls are excluded.

Here is an example of a suitable 8x8 input maze image:

Note that the closer you get to the walls when cropping, the better. No need to be exact though nor use any complex image manipulation tool. Taking a screenshot of the area with some care should be enough.

If you use GNU/Linux you might be able to take area screenshots by simply pressing MAYUS + PRINT and then selecting the area on your screen. Screenshots should appear on your home "Pictures" folder.

Command line options

You can list the command line options with:

ommr --help

The --size option specifies the size of the maze (i.e.: 16, 32...).

The --output specifies where to store the generated file:

• If ommited, the output file will be created in the same directory as the input file, using the input file name but changing the file extension to .txt.
• If the output is a directory, the same input file name will be used but the file will be created inside the specified output directory.
• If the output is not a directory it will be used as the output file name.