BlackJack Bot - Machine Learning, OpenCV

python   bots   OpenCV   machine learning   Qlearning   automation  

Created as a proof of concept for the more advance  Poker Bot. The aim was to solve  a number  of uncertainties  such as reading the screen using computer vision (OpenCV), simulating user input and using machine learning for the game logic.

The game logic was achieved using a custom created Machine Learning  (Q Learning) library that trained itself over several million simulated games. Black jack was selected for this as its a game using cards but has a much simpler game logic then poker.

Open CV in python was used to read the screen. Dealer and  player cards were detected along with the possible actions. 

This  was dreamed  a reasonably successful proof of concept with all the objectives achieved - and work was started on the ultimate heads up poker bot.