-
Notifications
You must be signed in to change notification settings - Fork 0
hell0w0rld2/Reconstruction-Generator
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
How to run: 1. Run calibrate.py. 2. Enter the name of a video and press enter. 3. From there, in the video, click the logo of the Rubik's cube first, then the corners and/or edges as if reading a book. 4. Copy output of calibrate.py, paste underneath TODO in opticalFlowTest.py 5. Run opticalFlowTest.py 6. Some notes on the two files: Red dot represents the 'calibrating point', should always stay on the logo of the cube Green dots track the movement of the corners and edges Press 'r' to reset the green points, to be done manually in between moves Press 'p' to pause Press 'q' to quit If red dot loses the logo, click on the logo once again to regain the center moveTracker.py: Called when the 'm' key is pressed. Determines the move performed using the changes in the location of points videos: solvedCube.mp4: Performing slow, outer layer turns on a solved cube. Used to enhance the optical flow map scrambledCubex.mp4: Performing slow, outer layer turns on a scrambled cube. Again, used to enhance the optical flow map and calculate point movements for each move. Multiple used videos to get a good average. scrambled notes: Stored version of each test run in opticalFlowTest.py. These files store the changes for each move to get a good idea of range and average change of each point to best recognize moves. Analyzed the data in scrambled notes by hard, as I found it faster than reformatting and collecting all of it in code
About
Using OpenCV's optical flow, I created a proof of concept for a rubik's cube reconstruction generator. All instructions for use are in the README.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published