Research On Active Visual Positioning Method Of Indoor Mobile Robot Based On Characteristics Of Points

Visual positioning technology has become a hot research topic in the field of robot.And it is also the difficulty in the field of robotics research.Although many scholars at home and abroad the new idea,the new method is proposed to implement the robot vision orientation,it is undeniable that the robot vision orientation in the study,there are still many problems to be solved.According to the robot visual locating methods are faced with the problem,put forward the corresponding solutions.This paper to do specific work is as follows:(1)The existing visual positioning method based on point feature,classification is introduced based on some theoretical foundation and research status of feature location method.Research situation including: feature points not at the same time,the effect on the positioning result;Iterative algorithm are summarized based on the characteristics of the positioned in the result the important effect of optimization.(2)The camera intrinsic parameters for five parameter model under the premise of design mathematical model of the robot vision system.With the original model used in four different parameter model,five parameter model considering the camera on the transverse and longitudinal directions method coefficient g,the coupling effects after model establishment and the original model in some details are different.Each of the parameters in the system were analyzed,and prepare for yuntai controller is designed in the fourth chapter.(3)According to the third chapter on the analysis of image jacobian matrix,the fourth chapter design with two degrees of freedom are the haeundae controller.The role of the controller is real-time adjustment yuntai pitching Angle and yaw Angle of the camera optical axis always aim at the reference target,the image feature information of ensure that appeared in the camera.Then through the analysis of the stability of the controller,the control algorithm can guarantee the system convergence,thus solve the vision constraint problems.(4)In the fifth chapter of the traditional visual locating method based on the four feature points made improvements,the improved algorithm,camera intrinsic parameters using the same five models.Visual positioning method based on the four feature points outside the parameters of the result is the camera to the world coordinate system,the parameters include the rotation matrix and translation vector.Pose the result accuracy is not high,this method obtained the results as a recursive least squares method of initial value,through recursive optimal positioning results,improve the positioning accuracy.Is outside the camera parameters obtained above,using the known,translation external parameters of camera and robot yuntai information feedback point of view,robot pose information is obtained.(5)The control method of the four rotor aircraft is studied in chapter six,the characteristics of four rotor position information and location information in the application of four rotor flight control.Comparing the original use of gyro accelerometer,GPS,air pressure meter,ultrasonic,laser equipment,such as computing position attitude information,the method of visual positioning technology can be a solution to complete the calculation of position and attitude information.Only a visual device can solve location attitude measuring task,saved the four rotor aircraft link of peripherals,more simple in hardware design.