Automatic Tracking of Surface targets with Vision Sensor

• Main Authors: Wu, Liang-Jye, 吳良杰
• Other Authors: Chen Chwan-Hsen
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/41780656737920564505

碩士 === 元智工學院 === 機械工程學系 === 84 === The visual tracking control system can be used in surveillance motion of an unmanned driving plane and in the monitoring system on boats to track surface objects. Also, it can be utilized in automatic driving of an unmanned vehicle and in robot''s assembly operation on a moving conveyor. To attain tracking control, it must be done as follow. First, the vision system needs to recognize a moving object from background. Second, the visual servo control must keep the object motionless in the image plane. To idenify the moving object, we detect the object through a series consecutive images acquired by a camera and estimate the position and velocity of the moving object. Then we control a servo motor to align the camera with the interested object in the picture frame. As the object is fixed in the frame, we can enlarge the image to recognize the shape of the target.This context presents an integrated control structure including a visual sensory subsystem of image processing and a servo control subsystem. The image ocessing subsystem uses the windowing and the subsample method to accelerate the processing speed. Meanwhile the servo control subsystem uses an image-based feedback control to manipulate the motion of the camera to keep the object in the center of image plane avoiding the task of finding the object position in the world coordinate and the environment calibration procedure. From our experiment setup, in which a camera tracks a swinging ping-pong ball, the sy stem response can attain near real time speed. Experiment results show automatic tracking of a surface target with a vision sensor is feasible and effective.