竞赛用投篮自动机器人目标识别方法研究

Research on Target Recognition Method of Automatic Shooting Robot for Competition

针对机器人在投篮过程中因视觉镜头转动幅度小、场地大,导致的识别精度低问题,提出一种考虑场地和空间因素的目标识别方法。分析竞赛用机器人的赛点规则,得出其在比赛中需要重点识别篮球架和罚球点的位置,采用色彩标记法篮球框、门柱、篮球架顶点、侧方点以及四边罚球点,计算每个点权重,通过形态学规则求得边缘特征、线性特征以及中心特征值。建立三维坐标系,计算摄像点与模拟目标点间映射关系,通过映射值分别求得目标点与篮球框在坐标轴三个方向上的角度及目标点与机器人在三方向的距离值,得到角度与距离阈值,根据现场参数代入识别实际目标点。仿真实验证明,所提方法识别精准度高、误差小,鲁棒性好,实用价值高。

Aiming at the problem of low recognition accuracy caused by small rotation range of visual lens and large field during shooting,a target recognition method considering field and space factors was proposed.Robots are used for analyzing the compe-tition rules of the match point,it is concluded that the need to focus on identifying and backboard in the game and the position of the spot,using color notation basketball frame,post,and backboard vertices,lateral and quadrilateral spot,calculating weights of each point,edge features,linear rules obtained through morphology characteristic value as well as the center.The 3d coordinate system was established to calculate the mapping relationship between the camera point and the simulated target point.The angles between the target point and the basketball frame in the three directions of the coordinate axis and the distance be-tween the target point and the robot in the three directions were obtained by the mapping values,and the Angle and distance thresholds were obtained.The actual target points were identified according to the field parameters.Simulation results show that the proposed method has high recognition accuracy,small error,good robustness and high practical value.