Rotation Detection And Tracking Of Mobile Robot Based On Visual Information Processing

Now with the development of robot technology,robot basic can undertake high difficulty action,like reversing course at any time,real-time detection of target,post rotation,etc.But,in the case of uneven ground or sandy soil,appear easily robot detector and sensor desensitization,so will be unable to accurately measure the rotation capacity.For these phenomena,the author will use visual information processing ways to further explore the rotation of the robot behavior and the ability to track targets.Install cameras in the robot,through camera to collect the corresponding information,to borrow the communication network will collect information transmitted to the robot internal server;Then the behavior person should decide to track the target,the selected data transmitted to the robot at the core of the control system;Core control system quickly for the rotation Angle and target distance is calculated,and then after the conversion of data transmitted to and thermal motion control system,the robot can act accordingly.This paper mainly discusses the two aspects of content: the first is the robot rotating ability is discussed in this paper.First when choosing a camera,in order to prevent the occurrence of image deformation situation,this article choose the black and white staggered grid to a fixed camera.Then,when the robot rotate,there will be a standard drawing and rotate the figure,by defining the distance length to determine the fixed point of the two figure,by measuring the Angle between the two fixed point again,can detect the rotation Angle of the robot.This article explores the second is the robot's ability to track targets.Detection technology was used in the test with TLD science,on the one hand,tested the robot hidden in the target and the target motion tracking ability,finally got good results;Control the robot,on the other hand,a summarization on the surface of the vehicle to track at random,the end result is that the robot is able to determine accurately track the target table and complete the required.Through a variety of experimental finally gained some good results:First,the detection of the robot when the rotation Angle,wholly in the SIFT science and technology.Through the inspection found that the robot to plane move to rotate or move first rotate again,gained by the rotation Angle are the same,thus come to a conclusion: plane does not move and rotate behavior order cause the change of the rotation Angle;To SIFT through the further optimization of science and technology,made the robot control system run speed increase 10 times faster;Through the robots to detect the rotation Angle and the rotation time data such as easy to calculate the rotating angular velocity.Second,the ability to track targets in detecting robot wholly in the TLD science and technology,and through regional division,such as sample to streamline method to optimize the TLD science and technology,eventually making the control system of speed running on 2 times,in accordance with the original scientific method processing every frame image takes 70 ms,but using the optimized only spend 30 ms science and technology.Through the balanced image precision and time,finally adopted 320 * 240 tracking image area.