Research On Object Identification And Tracking-Carrying For Mobile Robot

The contribution about visual servo is one of the research topics of mobile robot. The mobile robot tracking algorithm based on the monocular vision is discussed; overall framework to track and carry moving object is designed; based on the pan-tilt camera, a pure color object identification and tracking- carrying system is established.Firstly, visual scale-distance theory is expounded, and Zhengyou Zhang's camera calibration method is introduced, and this method is used to calibrate the pan-tilt camera. Some experiment results are discussed error precision of the visual scale-distance, and via dynamic calibration solves the position variety of the pan-tilt camera.For the pan-tilt camera is more smoothly and quickly to track the object, Kalman filtering algorithm is used to track moving object. Kalman filtering algorithm and its work process are introduced in this system.Secondly, from establishing the mathematical model of the camera, the relationship between the images target and the rotation angle of the pan-tilt can be described effectively, and introduced the motion of mobile robot and its carrying process.HSI and RGB color model is introduced. Some experiments are done to compare the both of color models. In addition, a seed filling algorithm is used to pick-up the target from the image information.Thirdly, under the VC + +. NET2003 environment, using VC + + and control function provided by mobile robot, a tracking-carrying system developed so that the control of pan-tilt and the robot's motion control function and convey modules are combined organically. The flow chart that presented and experimental results with their analysis are given at last.