TY - BOOK ID - 132616983 TI - Intelligent active vision systems for robots AU - Valdemar Cuevas Jimenez, Erik AU - Zaldivar Navarro, Daniel AU - Rojas, Raul PY - 2007 PB - Göttingen Cuvillier DB - UniCat KW - Artificial intelligence. Robotics. Simulation. Graphics KW - robots KW - computervisie UR - https://www.unicat.be/uniCat?func=search&query=sysid:132616983 AB - In this paper, an active vision system is developed which is based on image strategy. The image based control structure uses the optical flow algorithm for motion detection of an object in a visual scene. Because the optical flow is very sensitive to changes in illumination or to the quality of the video, it was necessary to use median filtering and erosion and dilatation morphological operations for the decrease of erroneous blobs residing in individual frames. Since the image coordinates of the object are subjected to noise, the Kalman filtering technique is adopted for robust estimation. A fuzzy controller based on the fuzzy condensed algorithm allows real time work for each captured frame. Finally, the proposed active vision system has been simulated in the development/simulation environment Matlab/Simulink. ER -