基于RoboMaster机器人运动目标的图像识别与处理

Image Recognition and Processing of Moving Target Based on RoboMaster Robot

以大学生创新实验室为平台,提出了一种对RoboMaster机器人装甲板的图像识别与跟踪方法.该方法首先依据RoboMaster机器人装甲板的光学特征和形状特征识别目标,利用其方位特征和运动特征跟踪目标,并使用PnP算法实时解算摄像头与目标装甲的物理距离,用于补偿云台射击目标时子弹由重力而出现的下坠,最后将转化后的目标位置信息传输给运动控制组.

This paper proposes a method for image recognition and tracking of RoboMaster robot armor plates by using the University Student Innovation Lab as a platform.The method first recognizes the target based on the optical and shape characteristics of the RoboMaster robot armor plate,tracks the target by using its orientation and motion characteristics and calculates the physical distance between the camera and the target armor in real time by using the PnP algorithm,which is used to compensate for the falling of bullets due to gravity when the pan-tilt shoots at the target,and finally transmits the converted target position information to the motion control group.