基于双目视觉和SIFT算法的羽毛球接球机器人避障研究

Research on obstacle avoidance of badminton ball catching robot based on binocular vision and SIFT algorithm

研究为了增强羽毛球接球机器人对羽毛球的识别和检测效率,并提高其接球效率和避障能力,提出了一种辅助羽毛球训练的接球机器人。首先利用无锚框检测的回归算法来实现机器人的目标检测和识别,此外引入了羽毛球轮廓识别技术和机器人避障技术来实现接球机器人的整体设计。最后通过尺度不变特征变换算法获取障碍物的环境特征点信息,使用加权最小二乘拟合法来获取障碍物的距离信息。在实验分析中显示,研究所提出的目标识别技术具有较低的误差,并长期保持在0.3以内,同时在羽毛球机器人的接球应用测试中显示,其避障反应时间较短,同时避障能力较强。特征点的误差整体控制在1~3的范围内。以上结果表明,通过双目视觉技术来实现羽毛球接球机器人的设计具有合理性,能够实现羽毛球的自动识别和接取,并有较强的避障能力,在羽毛球运动的智能化训练发展中具有重要价值。

In order to enhance the recognition and detection efficiency of badminton ball catching robot and improve its catching efficiency and obstacle avoidance ability,a kind of ball catching robot is proposed to assist badminton training.Firstly,the regression algorithm without anchor frame detection is used to realize the target detection and recognition of the robot.In addition,the badminton contour recognition technology and robot obstacle avoidance technology are introduced to realize the overall design of the ball catching robot.Finally,the environment feature point information of obstacles is obtained by scale invariant feature transformation algorithm,and the distance information of obstacles is obtained by weighted least squares fitting method.The experimental analysis shows that the proposed target recognition technology has a low error and remains within 0.3 for a long time.At the same time,the application test of badminton robot shows that its obstacle avoidance reaction time is short and its obstacle avoidance ability is strong.The error of feature points is controlled within the range of 1~3.The above results show that the design of badminton ball catching robot based on binocular vision technology is reasonable,can realize the automatic recognition and acquisition of badminton,and has strong obstacle avoidance ability,which has important value in the development of intelligent training of badminton.