Research And Design On Vision System Of The Mobile Robot

Mobile Robotics is one of the most important parts of Robotics. The research on Mobile robot started early in 1950's.With the development of the technology of Mobile robot, Robot is playing an increasingly important role in people's life. The robot faces in practical environment is the interaction between the robot and the environment. For the robot, efficient and robust vision system is of great importance for the control of the robot because the vision sensors can obtain most of the information about the environment. In this thesis, the research on the algorithms for mobile robots' vision and the development of control software for mobile robot are presented.Firstly, the research on the mobile robot technology and the vision system of mobile robot are reviewed, and the background and structure of this thesis are introduced.Secondly, the task and the structure of the mobile robot are presented in this thesis. Based on task of mobile robots' vision, the mobile robot' vision system of hardware is presented.Thirdly, we choose the HSI color module comparing with other color modules. After a thorough investigation on the algorithms of image processing and the object-recognition problem itself, v/e present a novel but easy color training algorithm. The simulation results show that the proposed method is effective and recommendable.Fourthly, Kalman Filter algorithm for estimating position and motion parameters of moving target centroid from monocular image sequence is proposed. Based on the characteristic of mobile robots' vision, using object tracking algorithm and manipulator controlling algorithm, we realize the robot' task to take and place the object.Finally, the results are summarized and future work is addressed.