基于点云分割的移动机器人运动目标跟踪方法

Mobile robot moving target tracking method based on point cloud segmentation

针对移动机器人在跟踪运动目标时受障碍物的影响,导致运动目标特征失真、视觉跟踪失败的问题,提出了基于点云分割的移动机器人运动目标跟踪方法,点云分割可以将目标物体划分为不同的子区域,从而挖掘重点区域,实现单独处理的目的。通过建立靶标世界坐标系,确定跟踪物体在世界坐标系的空间区域,采用空间投影点云分割方法,将多变特征分割映射到世界坐标系里的点云,确定跟踪目标感兴趣区域;通过对距离函数加权,融合决策层的传感器信息,运用视觉伺服与声纳伺服混合伺服跟踪方法,在确定跟踪目标区域内,实现移动机器人运动目标跟踪,最后设计了对比实验以验证此次方法的有效性。实验结果表明:该方法的点云分割均方根误差小,采用角速度6(°)/s、8(°)/s跟踪时的距离精确度高、中心距离误差低;在有、无障碍情况下对运动目标的跟踪偏移量均较小,稳定性强,可满足不同情况下移动机器人的实际跟踪需求。

Aiming at the problem that the mobile robot is affected by obstacles when tracking moving targets,resulting in the distortion of moving target features and the failure of visual tracking,a mobile robot moving target tracking method based on point cloud segmentation was proposed.Point cloud segmentation can divide the target object into different sub regions,so as to mine the key regions and achieve the purpose of separate processing.By establishing the target world coordinate system,the spatial region of the tracking object in the world coordinate system was determined.The spatial projection point cloud segmentation method was used to map the multivariate feature segmentation to the point cloud in the world coordinate system to determine the region of interest of the tracking target;By weighting the distance function,fusing the sensor information of the decision-making level,and using the hybrid servo tracking method of visual servo and sonar servo,the mobile robot moving target tracking was realized in the determined tracking target area.A comparative experiment was designed to verify the effectiveness of this method.The experimental results show that this method has small root mean square error of point cloud segmentation,high distance accuracy and low center distance error when tracking with angular velocity of 6(°)/s and 8(°)/s;With and without obstacles,the tracking offset of moving targets is small and stable,which can meet the actual tracking needs of mobile robots in different situations.