ChessBot: An Intelligent Robotic Chess Companion

Technical Documentation

ChessBot: An Intelligent Robotic Companion For Playing Chess

The ChessBot project is designed to create an intelligent robotic companion for playing chess. The architecture consists of several key components:

1. Inverse Kinematics: This module handles the inverse kinematics for the robotic arm, allowing it to move to specific positions based on the chessboard coordinates.
2. Chessboard Detection: The detect.py module uses OpenCV to detect the chessboard and its positions.
3. Piece Detection: The saddle_points.py module calculates saddle points and gradients to detect the presence of chess pieces on the board.
4. Model Training: The train.ipynb module trains a Convolutional Neural Network (CNN) to classify tiles as good or bad.
5. Prediction: The utils.py module contains functions to predict the presence of chess pieces and classify images.
6. Main Execution: The main.py and main.ipynb files contain the main logic for executing the chessbot.

Setup & Installation

To set up and install the ChessBot project, follow these steps:

1. Clone the Repository:

   git clone https://github.com/yourusername/ChessBot-ChessBot-An-Intelligent-Robotic-Companion-For-Playing-Chess.git
   cd ChessBot-ChessBot-An-Intelligent-Robotic-Companion-For-Playing-Chess
   ``

2. Download Pre-trained Models: Ensure you have the pre-trained models (patch_model_v2.pt and model.pt) in the appropriate directory.

The project uses several Python modules and functions. Below is a brief overview of the key functions and their usage:

• brutesac.py:

  • timed(f): A decorator to measure the execution time of a function.
  • get_all_quads(tri): Generates all possible unique quads from Delaunay triangles.
  • count_hits(given_pts, x_offset, y_offset): Counts matching points in a unity grid with given offsets.

• detect.py:

  • detect_chessboard_and_positions(image_path): Detects the chessboard and its positions from an image.

• model.py:

  • CNNModel(filter_sizes): Defines a CNN model for classification tasks.

• saddle_points.py:

  • get_saddle(gray_img): Calculates saddle points and gradients of a grayscale image.
  • fast_nonmax_suppression(img, win=11): Performs fast non-maximum suppression using dilation.
  • clip_bounding_points(pts, img_shape, win_size=10): Clips points near the image boundaries.
  • get_final_saddle_points(img, win_size=10): Gets the final saddle points from an image.

• train.ipynb:

  • Contains the training script for the CNN model.

• utils.py:

  • pred_patches(patches): Predicts patches using the pre-trained model.
  • predict_fn(inputs): Predicts probabilities for tiles.
  • classify_image(frame, win_size=10, prob_threshold=0.5): Classifies an image based on saddle points.
  • predict_on_tiles(tiles, predict_fn): Predicts on a batch of tiles.
  • predict_on_image(pts, img, predict_fn, win_size=10): Predicts probabilities for points in an image.

Configuration

Configuration settings are typically managed through environment variables or configuration files. Ensure the following environment variables are set:

• STOCKFISH_PATH: Path to the Stockfish executable.
• MODEL_PATH: Path to the pre-trained model files (patch_model_v2.pt and model.pt).