Browsing by Autor "Alberto Romay"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item type: Item ,
    Robust vision system to illumination changes in a color-dependent task
    (2011) Andrés Espínola; Alberto Romay; Tatiana Baidyk; Ernst Kussul
    Most computer vision tasks are strongly sensitive to illumination changes. This is the case of RoboCup competitions, being color dependent tasks, they require a robust color-based segmentation method of the image for object recognition. It is difficult to achieve a constant illumination in a work environment that is subject to day light changes. This is why camera parameters calibration such as exposure time must be robust enough to reduce the impact in the output images. A result comparison on the segmentation of images is made in three different color spaces, RGB, HSV, and YCrCb. We researched their connection with changes in lighting condition and select one of them that have less error for those environments. We introduce a fuzzy calibration method of the camera exposure time parameter, by histograms correlation. The calibrations results give an approximate value of the correct parameter to be modified in the camera in order to avoid affecting the thresholds of segmentation in the color recognition task.
  • Thumbnail Image
    Item type: Item ,
    Vision based localization of humanoid robots by inverse pose-estimation using a small set of fixed landmark features
    (2011) Luis F Lupian; Alberto Romay; Andrés Espínola
    We propose a vision based methodology to solve the robot localization problem by estimating the pose of the robot's camera with respect to a model of the environment represented by a small set of landmark features. The methodology proposed constructs a regression model based on the pinhole camera projection model by combining the information of several observed features. In this regression model the unknowns are the X and Y coordinates of the location vector of the robot with respect to the environment, and the depths of each of the observed features. We use a RoboCup Humanoid KidSize field as a test bed for experimental validation of this methodology.