Research on Robot Golfing Method Based on Feature Fusion and Kalman Filter

In order to reduce the recognition complexity and improve the detection accuracy and tracking efficiency, so that the robot can quickly and efficiently hit a golf ball into the hole, this paper proposes a method of combining the MEC + CEM detection algorithm and the Kalman filter assistant tracking algorithm. First, the acquired RGB image is transformed into HSV color space and the optimal threshold is taken. Then, the MEC + CEM detection algorithm is used to obtain the exact position of the golf ball. The matching information is used to correct the position information and the credibility judgment. Finally, the mathematical model of robot hitting is established. The experimental results show that the average accuracy of the algorithm can reach 94.06% for the detection of golf ball, which is better than the external disturbance factor and much faster than the Hough transform algorithm. In the success rate of hitting the hole, the robot using this algorithm is higher than human and meet the requirements of the golf tournament.