基于视觉目标跟踪的侦察机器人导航方法

Navigation method for reconnaissance robot based on vision object tracking

针对未知环境下侦察机器人的自主导航问题,提出了一种基于视觉目标跟踪的侦察机器人导航方法.首先利用二进制鲁棒独立元素特征(BRIEF)提取方法来检测和描述待跟踪视觉目标的局部不变特征点,在快速的特征匹配计算基础上提出由粗到精的目标定位两步法实现机器人导航过程中视觉目标的实时准确跟踪.其次对基于视觉目标跟踪的自主导航任务进行行为分解和实现,在行为中集成视觉目标跟踪算法.最后利用基于宏行为的机器人事务执行机制实现移向视觉目标的自主导航控制.实验结果表明,提出的方法能够使侦察机器人实时准确地跟踪视觉引导目标,在复杂障碍物环境下可靠地完成移向目标的自主导航任务.

Aiming at the problem of autonomous navigation of the reconnaissance robot in unknown environments, a navigation method for reconnaissance robots based on vision object tracking is pro- posed. First, the binary robust independent elementary features （BRIEF） extraction algorithm is em- ployed to detect and describe the local invariant feature points in the image region of vision object to track. Based on a fast feature matching procedure, a coarse-to-fine target location method is pro- posed for real-time and accurate object tracking in the robot navigation process. Secondly, the auton- omous navigation mission based on object tracking is decomposed into several behaviors in some of which the object tracking algorithm is integrated. Finally, the macro behavior-based robot mission execution mechanism is used to control the robot moving towards the vision target autonomously. Experimental results show that the proposed approach makes the reconnaissance robot accurately track the vision guidance object in real-time, and perform the target oriented autonomous navigation task reliably.