An Algorithm for Object Detection in Strategy Game Maps as An Alternative to Machine Learning Methods

Computer vision is perceived as a key sub-field of both artificial intelligence and machine learning. Application areas for their methods and software libraries have covered a very wide spectrum of issues: from handwriting recognition, via analysing real-time images and intelligent automation to games and entertainment. The paper introduces an image recognition algorithm for detection of hexagon fields and environmental objects (like roads, rivers, lakes, forests, mountains, swamps and cities) within maps designed for strategy board games. The algorithm can be used to ease the process of migration of these board games into systems supporting such gameplays. Due to sustained Covid-19 restrictions and thus growing popularity of board games it has been necessary to allow ones to enjoy these games online and making the games available to a wider audience. The presented approach uses OpenCV and the dedicated methods for object detection in multiply colour spaces.