Crowns / Queens Puzzle Solver

A computer-vision and constraint-based solver for Crowns (Queens) logic puzzles.

The system takes a puzzle screenshot as input, reconstructs the grid and regions using OpenCV, solves the puzzle via constraint logic, and re-renders the board from scratch with queen placements using Canvas.

Key Features

• 📸 Screenshot-based puzzle input
• 🔍 Automatic grid detection and cell segmentation
• 🎨 Region detection via color clustering
• 🧩 Constraint solver for queen placement
• 🎭 Clean board re-rendering (Canvas)
• 🏗️ Clear separation of vision, solver, and UI layers

How It Works

1. Frontend uploads a puzzle screenshot
2. Backend extracts the grid using OpenCV
3. Cells are grouped into regions by color
4. Solver applies constraint rules
5. UI renders a clean puzzle board and queen positions

The screenshot is used only as input.

Tech Stack

Backend

• Python
• OpenCV
• NumPy
• FastAPI

Frontend

• HTML
• JavaScript
• Canvas API

Quick Start

1. Install dependencies:

pip install -r requirements.txt

2. Run the web app:

cd WebApp/
uvicorn main:app --reload 

The application will be available in your browser. Upload a puzzle screenshot and click Solve.

3. (Optional) Run tests:

python Tester.py

API Output (Simplified)

{
  "cells": [
    [{ "region": 0, "color": [214, 181, 92] }]
  ],
  "queens": [
    { "row": 0, "col": 1, "region": "C1" }
  ]
}

• cells → UI-ready color grid
• queens → solved queen positions

Project Structure

├── CoreSolver.py              # N-Queens constraint solver
├── ImageProcessing.py         # Grid detection & cell segmentation
├── GameofCrownsBoardFetcher.py # Board data fetching
├── Tester.py                  # Test suite
├── requirements.txt           # Python dependencies
├── images/                    # Sample puzzle screenshots
└── WebApp/
    ├── main.py                # Flask/FastAPI backend
    ├── index.html             # Web UI
    ├── app.js                 # Canvas rendering + API calls
    └── __init__.py

Design Highlights

• No screenshot overlays → no alignment bugs
• Separate color pipelines for solver and UI
• Deterministic rendering
• Canvas-based drawing for precision

Output

Limitations

• Requires a clear grid in the screenshot
• Similar region colors may need threshold tuning
• Square grids only (for now)

Roadmap

• Step-by-step solver visualization
• Adding DFS to core solver as additional mechanism
• Export solved boards as images

License

MIT