基于单目视觉的无人水面艇目标跟踪方法

Monocular vision-based method for target tracking of USV

水面目标跟踪是无人水面艇(USV)实现智能作业的重要基础技术。针对流场干扰下传统船舶纯方位目标跟踪存在较大跟踪误差的问题,提出一种基于单目视觉的无人水面艇目标跟踪方法。在构建跟踪船与被跟踪船运动模型的基础上,利用跟踪船的单目相机获取被跟踪船船尾搭载的标志物图像。通过比较运动过程中标志物图像与期望图像的位置关系,实时计算出偏航角和偏航距离作为控制器输入量,实时调节跟踪船的航向和航速,从而实现对被跟踪船的跟踪。仿真实验分析了不同目标路径情况下流场干扰对跟踪性能的影响,实验结果表明:该无人水面艇目标跟踪系统具有良好的跟踪性能和一定的抗干扰能力。

Surface target tracking is one of the fundamental technologies for unmanned surface vehicle(USV)to realize intelligent operation.Aiming at solving the problem of large tracking errors in azimuth-only target tracking of traditional ships under flow field interferences,a target tracking method based on monocular vision is proposed for USV.Based on the motion model of the tracking ship and the tracked ship,the image of the marker on the stern of the tracked ship is obtained by the monocular camera of the tracking ship.By comparing the position relationship between the marker image and the desired image during the movement,the heading error and cross-track error are calculated in real time and are used as the controller inputs,and the heading and speed of the tracking ship are adjusted accordingly to realize the tracking of the tracked ship.By simulation experiment,the influence of flow field interferences on tracking performance for different target paths are analyzed.The results show that the target tracking system of USV has good tracking performance and certain interference rejecting properties.