Analyze an image, directory of images, or video feed to locate a rubiks cube. The RGB values for each square will be printed to STDOUT.

This works for 2x2x2, 3x3x3, 4x4x4, 5x5x5, and 6x6x6 cubes. 6x6x6 is the largest cube I have test with but 7x7x7 and larger should also work if all of the squares are the same size.

If you are using the --webcam option a solution will be displayed on the screen for 2x2x2, 3x3x3 and 4x4x4 cubes. I have a solver for 5x5x5 but it takes about a minute to compute the solution and the solutions tend to be 90+ steps so this doesn't work very well for displaying on camera.

$ sudo apt-get install python-pip python-opencv
$ sudo pip install git+https://github.com/dwalton76/rubiks-cube-tracker.git

Follow the instructions at https://github.com/dwalton76/rubiks-color-resolver to install the rubiks-color-resolver library

Follow the instructions at https://github.com/dwalton76/rubiks-cube-NxNxN-solver to install a solver for 2x2x2, 3x3x3, 4x4x4, 5x5x5, 6x6x6, and 7x7x7 cubes

$ rubiks-cube-tracker.py --webcam 0

• --webcam 0 means use /dev/video0, --webcam 1 means use /dev/video1, etc
• Press the SPACEBAR to scan a side
• You MUST scan the sides in F R B L U D order. Once you've scanned L you want to put F back in front and then flip forward one time to scan U, this way U is oriented correctly. To go from U to D just flip forward two times. Once you've scanned D, D will be facing the camera, U will be facing you and F will be on top.
• Flip the cube so that U is on top and F is facing you, then follow the solve steps on the screen.
• Press "r" to reset

Analyze a single image. This is only used for debugging and will pop up images at various stages of locating the squares.

$ rubiks-cube-tracker.py --filename test/test-data/3x3x3-random-01/rubiks-side-B.png

Analyze a directory of images where the files are named rubiks-side-U.png, rubiks-side-L.png, etc

$ rubiks-cube-tracker.py --directory test/test-data/3x3x3-random-01/