Solving sudoku Realtime with the help of Computer Vision , Backtracking Algorithm and Artifictial Intelligence
Sudoku is one of the most popular puzzles. The goal of sudoku is to place the numbers one through nine exactly once in each row, column, and 9x9 box based on the starting numbers that are already filled in and any subsequent numbers you fill in. In this project, we aim to create a real-time sudoku solver which recognizes the sudoku grids and provide solution with the help of Computer Vision.
Python version 3.7.7 is used in this project.
OpenCV is a cross-platform library using which we can develop real-time computer vision applications. It mainly focuses on image processing, video capture and analysis including features like face detection and object detection.
To install opencv:
- In Linux:
pip install opencv-python - In Conda environment:
conda install -c conda-forge opencv
Python-tesseract is an optical character recognition (OCR) tool for python. That is, it will recognize and “read” the text embedded in images.Python-tesseract is a wrapper for Google’s Tesseract-OCR Engine.
To install opencv:
- In Linux:
pip install pytesseract - In conda environment:
conda install -c conda-forge tesseract
Python-tesseract requires path to be set, if you are like me who couldnt set the path during installation. Then find where your tesseract.exe file is and then copy that path and paste it on webcam_main.py file pytesseract.pytesseract.tesseract_cmd = r'yourpath' instead of pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe' . If you have set the path prior then comment out the pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe' line in webcam_main.py file.
To run this project, follow the steps given below:
Clone the repositoryhttps://github.com/Anjitha95/Real-time-sudoku.git
Run the file webcam_main.py
The image is converted to grayscale and Adaptive Thresholding are applied to the image to reduce noise. Then the maximun area using hough lines and contouring is found which is assumed to be sudoku grid.
By the cordinates we found using contour and houghlines, we wrap the grid and extracts out the individual cells out of the sudoku grid.
Here the individual cell images are passed to Python-tesseract (OCR) to recognize the image and to return the set of recoginzed image as string.which is then converted to a 2D array
The array of recognised image is feed to a backtracking algorithm to solve the sudoku and to print out the sudoku solution.
The solution is then overlayed on the empty sudoku grid image
- https://stackoverflow.com/
- https://towardsdatascience.com/open-cv-based-sudoku-solver-powered-by-rust-df256653d5b3
- https://medium.com/@pmprakhargenius/sudoku-solver-ai-agent-700897b936c7
- https://www.youtube.com/
- https://github.com/manpreet1130/RealTime-Sudoku-Solver
And many more........
- There exists a lag while capturing and contoring the image
- When sudoku is not solvable it is gives back the same unsolved array as the result.
- The solver cannot understand whether the shown grid is already solved or not, so it solves each time a sudoku is shown.
- Augmented reality