Detection of a Spot of Pre-defined Color Using a Video Color System in the Context of Robotics

Authors: Sadettin KAPUCU, Sedat BAYSEÇ

Abstract: In robotics, vision provides the control system with vast amounts of data concerning the manipulator itself and its environment. Many different image identification techniques have been proposed since the introduction of vision systems to robotics, but as the size of data is large, most of these techniques require so long computation times, not applicable at shop floor. This paper suggests the use of specific colors which are unique to the object of interest enabling their detection and hence distinction from the rest of the environment easily, irregardless of the details and intricacy they may bear. Such brightly colored spots can be put on an object in precise locations during manufacturing. The problem then comes up as to search the color definitions of the pixels forming the image and select the ones which fit to the pre-defined color. In this paper, to define the ability of distinguishing colors, Colorimetric System, Nevatia Transformations and NTSC Color Systems are tested and results are presented.

Keywords: Computer Vision, Colorimetric System, Nevatia Transformations, NTSC Color System.