Research And Design Of Human Tracking Robot

Human recognition and tracking is an important research direction in intelligent videosurveillance and computer vision, and is a basic technology of service robots. Aiming at thehuman recognition and tracking in family service robot, this paper proposes multi-informationfusion of human body tracking algorithm and speaker-dependent recognition algorithm. Ahome service robot is designed, applying this algorithm to service robots, the tests of thealgorithm show that it can realize robust tracking and identification of human body incomplex background, similar color and the changes of light.Firstly, multi-information fusion of human target tracking algorithm is proposed forhuman recognition and tracking, Camshift algorithm is expended into three-dimensionalspace by color information and depth information. Then integrating depth information of thehuman body target in three dimensional space, predictiing target position information andeliminating the influence of complex background using Kalman filter. At last, human object’slocation information is forecasted by Kalman algorithm and human’s center of mass locationis calculated in the corresponding depth image, human object’s tracking is achieved now.Using Kinect sensor to realize multi-information fusion of human target tracking algorithm,the experimental result shows that the algorithm proposed in this paper can track motionhuman robustly in condition of complex background, uneven ambient light, deformed targetand so on.Secondly,this paper proposes a speaker-dependent recognition algorithm based on thedistribution of color features, the algorithm firstly selected the shoulder width as the length,the rectangle of right shoulder to the human body center of gravity height as color histogramstatistics, identify the target using histogram cross matching method. Test results show thatthe algorithm can quickly and accurately find the owner and follow even if the human bodytarget tracking, and is not sensitive to light and changes of environment.Finally, human body target detection and tracking system was designed and implementedin the service robot platform, this paper uses the Kinect sensor, completes the overall systemand each function module design using C and C++mixed programming, And a service robotis designed,we show the detail composition, function and hardware structure.At last we usethe service robot track in practical scenario, the test results show that the robot can not onlyrobust tracking human body target, and can search the master and follow up when the ownerlost.