Color-spaces and color segmentation for real-time object recognition in robotic applications

Abstract

Object recognition plays an important role in a vast diversity of robotic applications. This paper1 addresses the vantages and disadvantages of using some known color spaces (digital color representations) and segmentation methods for the purpose of accurate detection of color-coded objects, which are objects that are represented by a given color. The robots are previously programmed to associate each color or combination of colors to the type of object to be detected. The subject of shadow suppression based solely on color manipulation will also be addressed.