sudoku-solver

This program allows you to recognize and solve Sudoku puzzles using OCR and a Sudoku solver written in C++.

Prerequisites

Before using this program, ensure that you have the following dependencies installed:

Then install the required Python packages by running the following command:

pip install -r requirements.txt

Place the screenshot of the Sudoku puzzle in a file named screenshot.png. Ensure that the puzzle is clear and visible in the image.
Run the OCR program by executing the following command:
```
python scan.py
```
The OCR program will process the image, recognize the numbers, and save the OCR result in a file named sudoku.txt. Alternatively, you can put the puzzle directly into sudoku.txt. The file sudoku-example.txt shows what a typical puzzle would look like.
Build the C++ Sudoku solver program using CMake. In the project directory, execute the following commands:
```
cmake .
cmake --build .
```
This will generate the necessary build files and compile the C++ program.
Run the Sudoku solver program by executing the generated executable. Use one of the following commands:
```
./sudoku 10  # find 10 solutions (same as `./sudoku`)
./sudoku alt # find the solution step by step (might fail for some puzzles)
```
The program will read the Sudoku puzzle from the sudoku.txt file, solve it, and display the solutions.

Ensure that the Sudoku puzzle in the screenshot is clear and well-aligned. Blurry or distorted images may affect the OCR accuracy.
If the OCR results are not accurate, you can experiment with different image preprocessing techniques, such as adjusting brightness/contrast, applying filters, or resizing the image.
Make sure that the Sudoku puzzle is a standard 9x9 grid with numbers from 1 to 9. The OCR program and Sudoku solver are designed to work with standard Sudoku puzzles.
Feel free to modify and extend the OCR program and Sudoku solver to suit your specific needs or add additional features.
This README is mainly written by ChatGPT.

Happy Sudoku solving!