摘要
合作靶标点三维轨迹的跟踪识别是实现室内环境中多飞行器位姿估计的关键,为此,提出了一种基于时空一致条件下的多目标三维轨迹跟踪识别算法。该方法包括运动轨迹跟踪与识别两部分,对于合作靶标点三维轨迹跟踪,提出了一种基于运动目标位移矢量一致的数据关联方法,该方法首先利用运动平滑性假设计算得到的数据关联概率值,结合匈牙利算法求解得到目标的数据关联关系,然后在贝叶斯滤波框架下实现合作靶标点的三维轨迹跟踪。对于合作靶标点的三维轨迹识别,又可以分为粗细两部分,利用运动轨迹Hankel矩阵的秩实现运动轨迹的粗识别,利用运动轨迹之间的Hausdorff距离实现运动轨迹的细识别,最终实现对每一个飞行器的轨迹识别与注册。实验结果表明,在三维测量手段为机器视觉,测量空间大小为2 m×2 m×2 m,提出的多目标跟踪算法的三维轨迹跟踪误差小于4 mm(3σ)时,轨迹识别正确率为100%。因此,所提出的算法可以有效地实现多飞行器上合作靶标点三维轨迹的跟踪识别。
The tracking and recognition of the three-dimensional trajectories of cooperative target points is the key to the estimation of the position and attitude of multi-aircraft in an indoor environment. Therefore, a multitarget three-dimensional trajectory tracking and recognition algorithm based on time and space consistent conditions was proposed. This method included two parts of motion trajectory tracking and recognition. For the three-dimensional trajectory tracking of cooperative target points, a data association method based on the consistency of the displacement vector of the moving target was proposed. This method first used the data association probability calculated by the motion smoothness assumption. Combined with the Hungarian algorithm to solve the target data association relationship, and then the three-dimensional trajectory tracking of cooperative target points under the Bayesian filter framework was realized. The three-dimensional trajectory recognition of cooperative target points was divided into two parts: the rank of the motion trajectory Hankel matrix to realize the coarse recognition of the motion trajectory, and the Hausdorff distance between the motion trajectories to realize the fine recognition of the motion trajectory. Eventually the trajectory recognition and registration of each aircraft was realized. Under the experimental conditions of computer vision measurement method and 2 m×2 m×2 m measurement space, the results show that the proposed multi-target tracking algorithm has a three-dimensional trajectory tracking error of less than 4 mm(3σ), and the trajectory recognition accuracy rate is 100%. Therefore,the proposed algorithm can effectively realize the tracking and recognition of the three-dimensional trajectory of cooperative target points on multi-aircraft.
作者
霍炬
何明轩
李云辉
薛牧遥
Huo Ju;He Mingxuan;Li Yunhui;Xue Muyao(Department of Electrical Engineering,Harbin Institute of Technology,Harbin 150001,China;School of Mechatronic Engineering and Automation,Shanghai University,Shanghai 200444,China;Space Propulsion Technology Research Institute,Shanghai Academy of Spaceflight Technology,Shanghai 201109,China)
出处
《红外与激光工程》
EI
CSCD
北大核心
2020年第10期225-233,共9页
Infrared and Laser Engineering
基金
国家自然科学基金(61473100,61021002)。
关键词
多目标跟踪
轨迹识别
数据关联
矢量一致
multi-target tracking
trajectory recognition
data association
vector consistency
