Invariant segmentation to the luminosity of robots by using Microsoft Kinect Sensor and Otsu's method CANTERO, S. V. A. B. ; GONÇALVES, W. N. XI Workshop de Visão Computacional, p. 376-381, 2015
With the evolution of Robotics is requested therobots automation and the use of techniques from the areas ofartificial intelligence and computer vision are essential. Footballmatches between robots is a very interesting scenario that allowsthe realization of experiments and real tests for perfecting therobots. The matches are a form inspiring in the study of robotics,computer vision and artificial intelligence, because only with acamera getting a bird's-eye view of the football field, you shouldimplement a strategy game that feeds of information generated byimage processing. This article proposes an efficient segmentationmethod of the robot invariant to brightness, interesting fact sinceother computer vision systems depend on a lot of light. MicrosoftKinect, which has an RGB camera, obtains the color imagesand an infrared camera returns the distance (depth) betweenMicrosoft Kinect and the objects captured by the camera. Thesystem consists of two steps, first the image of infrared camerais segmented through the depth and subsequently only the areaof interest is segmented again using the Otsu method on a RGBimage.