Font Size: a A A

Badminton Robot Target Detection And BEZIER Curve Path Planning

Posted on:2018-01-29Degree:MasterType:Thesis
Country:ChinaCandidate:Y ChangFull Text:PDF
GTID:2358330518452601Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
The research of badminton robots is mainly computer vision technology,target detection and tracking algorithm,image recognition,control engineering and so on.In this paper,we study and discuss the target recognition and motion control of the badminton robots from two aspects:Firstly,the overall structure of the badminton is introduced,including the ball-driving mechanism and the omni-directional mobile platform.The service mechanism is optimized.The acceleration,speed and equivalent quality of the shuttle are analyzed from the angle of service and the distance of the cylinder.Influence of Badminton Falling Range.Select the Kinect sensor as the vision of the badminton robot.First of all,the use of Zhang Zhengyou calibration Kinect for ordinary camera and depth camera calibration of-the internal parameters,improve the clarity of image acquisition.In the process of image preprocessing,the depth image is seriously noisy,and the weighted average method is used for gray scale processing.The image denoising is denoising using mean filtering,and the optimal 5 × 5 function window.Then the modified ICP algorithm is used to reconstruct the 3D surface of the badminton.As the badminton is not a closed object,the edge part of the badminton will be lost when modeling,but the whole effect can meet the requirements.In the process of target recognition and trajectory tracking,this paper adopts the latest kernel correlation filter tracking algorithm,which uses the algorithm to identify and track the badminton in a single background and complex background,the effect is good and effective Tracking accuracy.The design of omni-directional mobile platform,including the kinematics model,the dynamic model,the velocity and the velocity of the moving platform,and the dynamic model of the omni-directional mobile platform are analyzed comprehensively in the motion control of omnidirectional moving mechanism.Acceleration domain,and the DC brushless motor simulation model.The motion path of the robot is modeled by Bezier curve,and the speed of the three moving wheels under this path is simulated and the motion error is analyzed.The results show that both the velocity error and the motion error are feasible,and the path planning method is feasible.
Keywords/Search Tags:Badminton robot, Machine vision, Image processing, Brushless DC motor modeling, Motion control
PDF Full Text Request
Related items