Research Of Mobile Robot Visual Tracking Technology

In recent years,with the rapid development of the industrial robot and the great progress of artificial intelligence at home and abroad,the robots are playing an increasingly important role in human life.Meanwhile the state put forward the establishment of intelligent and safe cities by all things perception,interconnected and intelligence,that makes the national security industry obtain an opportunity of rapid development.On the basis of this,the machine vision has become the research hotspot and attracts much more attention.The robot can access large amounts of information through the visual system which just like human eyes.And then,it can make a final decision by processing the information that obtained before.In the current study,a mobile robot is selected as the research object.The study is mainly focus on how to make the tracking robot achieve to track the navigation robot at various movement pattern accurately and stably by designing a effective visual tracking algorithm.The tracking task of mobile robots can be accomplished by the visual tracking control system and robot motion control system.Three tasks have been completed experimentally based on reading a large number of literatures,showing as follows:(1)Firstly,the research destination,background and status of robot vision tracking control system are introduced in detailed.Then,the basic working principle of monocular vision is studied deeply,including the monocular camera calibration technique,monocular vision range finding principle and so on.Finally,series experiments have been carried out to verify the accuracy of the monocular vision range finding principle.(2)The research of the visual tracking algorithm: the Mean Shift algorithm and CamShift algorithm are analyzed deeply first.Then,the particle filter is introduced into the field of tracking control.In addition,an improved particle filter based on the CamShift tracking algorithm is proposed in this paper.Series of comparative experiments have been conducted at last to prove the effectiveness of this tracking algorithm and the results demonstrate that the accuracy and stability of visual target tracking have been improved.Therefore,the improved particle filter based on the CamShift tracking algorithm proposed in this paper can achieve to provide the mobile robot tracking control system with a stability visual tracking algorithm.(3)The realization of the mobile robot vision tracking control system: First,the tracking information has been obtained through the particle filter based on the CamShift tracking algorithm.The obtained information can be treated as the basis for the mobile robot motion and used to design visual tracking control strategy.Then,the change of the distance and target centroid between the tracking robot and navigation robot can be used as basis for judgment.Finally,a movement control strategy is designed to make the navigation robot always stays in the view of the tracking robot,on the basis of this,the mobile robot vision tracking control system can be realized.On the basis of OpenCV image processing library of Microsoft Visual C + + 6.0 software development environment,the Microsoft Visual C + + 6.0 in the Microsoft Foundation Classes is used to develop an interface interaction program which can make the tracking robot track the navigation robot successfully at the laboratory environment.