基于单目视觉伺服系统的高速旋转球体三维速度测定

发布时间：2019-06-20 14:55

【摘要】：单目视觉所处理的信息是目标物体在三维空间内的信息向二维平面的投影,故还原目标的深度信息,对于预测目标轨迹和后续控制决策具有重要的意义。提出一种基于单目视觉的高速旋转球体三维速度的方向及大小的测定算法,利用球体上单个标记的运动轨迹来识别球体旋转速度及旋转轴的方向。摄像机角度固定,当球体旋转轴正对摄像机时,球体标记运动形成的运动轨迹是一个圆,而当球体旋转轴方向发生改变时,球体标记运动轨迹将会变成一个椭圆,不同的方向形成椭圆的长短轴不一样,利用这个特征来计算旋转轴的方向。该算法可用于测量乒乓球旋转中,旨在应用于乒乓球机器人的视觉识别,对机器人后续阶段的运动规划和控制决策起到重要的作用。对所提算法进行了实验验证,得到与理论结果误差较小的实验数据,验证了方法的可行性。
[Abstract]:The information processed by monocular vision is the projection of the information of the target object to the two-dimensional plane in three-dimensional space, so it is of great significance to restore the depth information of the target for predicting the target trajectory and subsequent control decision. An algorithm for measuring the direction and size of three-dimensional velocity of a high-speed rotating sphere based on monocular vision is proposed. The motion trajectory of a single mark on the sphere is used to identify the rotation velocity of the sphere and the direction of the rotation axis. The angle of the camera is fixed. When the rotation axis of the sphere is aligned with the camera, the trajectory formed by the ball mark motion is a circle, while when the direction of the rotation axis of the sphere changes, the trajectory of the sphere mark movement will become an ellipse, and the long and short axis of the ellipse will be formed in different directions. The direction of the rotation axis will be calculated by using this feature. The algorithm can be used to measure the rotation of table tennis. It can be used in the visual recognition of table tennis robot and plays an important role in the motion planning and control decision of the robot in the subsequent stage. The proposed algorithm is verified by experiments, and the experimental data with small error with the theoretical results are obtained, and the feasibility of the method is verified.
【作者单位】：上海体育学院中国乒乓球学院;上海理工大学光电信息与计算机工程学院;
【基金】：上海市科学技术委员会科研计划项目(15490503100) 国家体育总局科研课题(2014B072) 上海体育学院研究生教育创新基金(yjscx2016007)
【分类号】：TP391.41

【相似文献】