Skip to content

Latest commit

 

History

History
41 lines (24 loc) · 1013 Bytes

README.md

File metadata and controls

41 lines (24 loc) · 1013 Bytes

Chessboard Movement Detection

Purpose

Creating a system that track the game and show the movements and records at the end of the match.

Techniques

Only image processing techniques were used to create the system.

There isn't any deep learning model or machine learning algorithm.

We can simply list the processing steps as below

  • Convert the board image into Bird'eye view
  • Check if hand is on the board
  • Find checker contours
  • Find the position of each checker on the board
  • Seperate white and black checkers
  • Compare previously saved state with the current state and find the differences
  • Save the changes (Player, Checkter Old Position, Checker New Position, Defeated Rival Checkers)

Visual Results

Original Video Frame:

alt text

Bird'eye View Frame:

alt text

When the player makes a move

alt text

alt text

When player defeat a rival checker

alt text