Analýza obrazu a detekce objektů pro hru fotbal robotů

Abstract

This thesis deals with a problem of analyzing image in robot soccer. We showcase few basic methods of image processing and we create new algorithm for detecting specific informations about the field and players. This is done by using simple edge detection method with thinning of the edges afterwards. The algorithm for image understanding is using features of individual objects to find out their whereabouts. The image processing runs mostly in realtime. Final product is new module for newly formed library for robot soccer.