摄像机网络中的一种多目标分布式跟踪策略
|
摘要
为了实现摄像机网络中的多目标分布式跟踪和解决存在的"朴素"问题,提出了一种基于信息一致性的多目标分布式跟踪策略。首先通过对分布式多目标跟踪问题的建模,得到基于平均一致性和信息加权一致性的网络中每个节点的最佳状态估计值及其信息矩阵;然后通过考虑虚假测量值和一致性融合,将来自于多个传感器的测量值与具体的目标关联起来,得到采用信息形式融合来自于多个传感器的测量值的状态估计方程,从而实现对多目标的数据关联和分布式跟踪。仿真实验结果表明,提出的多目标信息一致性跟踪算法相比于目前几种先进的跟踪算法,在平均误差性能方面不仅对于虚假测量值/杂波具有更好的鲁棒性,而且也有更好的收敛性。
In order to implement multi-target distributed tracking in camera network and solve the problem of "naivety",a multi-target distributed tracking strategy based on information consensus is proposed in this paper. Firstly,by modeling the distributed multi-target tracking problem,the optimal state estimation value and its information matrix for each node in the network are obtained based on average consensus and information weighted consensus. Then,by considering the false measurements and consensus fusion,the measurements from multiple sensors are associated with specific target,and the state estimation equations are obtained by using information form for combining measurements from multiple sensors,so as to implement the data association and distributed tracking for multiple targets. The simulation results show that the proposed multi-target information consensus tracking algorithm has better robustness and convergence than the other advanced tracking algorithms in terms of average error performance.
In order to implement multi-target distributed tracking in camera network and solve the problem of "naivety",a multi-target distributed tracking strategy based on information consensus is proposed in this paper. Firstly,by modeling the distributed multi-target tracking problem,the optimal state estimation value and its information matrix for each node in the network are obtained based on average consensus and information weighted consensus. Then,by considering the false measurements and consensus fusion,the measurements from multiple sensors are associated with specific target,and the state estimation equations are obtained by using information form for combining measurements from multiple sensors,so as to implement the data association and distributed tracking for multiple targets. The simulation results show that the proposed multi-target information consensus tracking algorithm has better robustness and convergence than the other advanced tracking algorithms in terms of average error performance.
引文
[1]ZHAO Y,DUAN Z,WEN G.Finite-time consensus for second-order multi-agent systems with saturated control protocols[J].Control Theory&Applications IET,2015,9(3):312-319.
[2]FANG W,WANG D H,WANG Y.Energy-efficient distributed target tracking in wireless video sensor networks[J].International Journal of Wireless Information Networks,2015,22(2):105-115.
[3]KAMAL A T,BAPPY J H,FARRELL J A,et al.Distributed multi-target tracking and data association in vision networks[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2016,38(7):1397-1410.
[4]RANA M M,LI L.Microgrid state estimation and control for smart grid and Internet of Things communication network[J].Electronics Letters,2015,51(2):149-151.
[5]SONG H,YU L,ZHANG W A.Distributed consensusbased Kalman filtering in sensor networks with quantised communications and random sensor failures[J].Signal Processing IET,2014,8(2):107-118.
[6]KAMAL A T,FARRELL J A,ROY-CHOWDHURY A K.Information weighted consensus filters and their application in distributed camera networks[J].IEEETransactions on Automatic Control,2013,58(12):3112-3125.
[7]董文会.多摄像机监控网络中的目标连续跟踪方法研究[D].济南:山东大学,2015.DONG W H.Research of target continuous tracking inmulti-camera surveillan networ[D].Jinan:Shandong University,2015.
[8]谭飞刚,黄玲,刘建.多线索融合的无视野重叠跨摄像机行人跟踪算法[J].计算机应用研究,2017,34(8):2542-2545.TAN F G,HUANG L,LIU J.Pedestrian tracking across nonoverlapping camera views based on fusion multi-clue[J].Application Research of Computers,2017,34(8):2542-2545.
[9]蒋建国,顾占冰,胡珍珍,等.多摄像机视域内的目标活动分析[J].电子学报,2014,42(2):306-311.JIANG J G,GUZ B,HU ZH ZH,et al.Activity analysis cross muti-camera[J].Acta Electronica Sinica,2014,42(2):306-311.
[10]MEYER F,BRACA P,WILLETT P,et al.A scalable algorithm for tracking an unknown number of targets using multiple sensors[J].IEEE Transactions on Signal Processing,2017,65(13):3478-3493.
[11]SUNG Y,CHUNG W.Hierarchical sample-based joint probabilistic data association filter for following human legs using a mobile robot in a cluttered environment[J].IEEE Transactions on Human-Machine Systems,2016,46(3):340-349.
[12]HAN J,TAN T,CHEN L,et al.Moving objects representation for object based surveillance video retrieval system[J].International Journal of Security&Its Applications,2014,8(2):315-322.
[13]BOUCHRIKA I,CARTER J N,NIXON M S.Towards automated visual surveillance using gait for identity recognition and tracking across multiple non-intersecting cameras[J].Multimedia Tools&Applications,2016,75(2):1201-1221.
[14]BRACA P,GOLDHAHN R,FERRI G,et al.Distributed information fusion in multistatic sensor networks for underwater surveillance[J].IEEE Sensors Journal,2016,16(11):4003-4014.
[15]SHANG L,ZHAO K,CAI Z,et al.An energy-efficient collaborative target tracking framework in distributed wireless sensor networks[J].International Journal of Distributed Sensor Networks,2014,2014(2):1-18.
[16]SHI K,CHEN H,LIN Y.Probabilistic coverage based sensor scheduling for target tracking sensor networks[J].Information Sciences,2015,292(C):95-110.
[2]FANG W,WANG D H,WANG Y.Energy-efficient distributed target tracking in wireless video sensor networks[J].International Journal of Wireless Information Networks,2015,22(2):105-115.
[3]KAMAL A T,BAPPY J H,FARRELL J A,et al.Distributed multi-target tracking and data association in vision networks[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2016,38(7):1397-1410.
[4]RANA M M,LI L.Microgrid state estimation and control for smart grid and Internet of Things communication network[J].Electronics Letters,2015,51(2):149-151.
[5]SONG H,YU L,ZHANG W A.Distributed consensusbased Kalman filtering in sensor networks with quantised communications and random sensor failures[J].Signal Processing IET,2014,8(2):107-118.
[6]KAMAL A T,FARRELL J A,ROY-CHOWDHURY A K.Information weighted consensus filters and their application in distributed camera networks[J].IEEETransactions on Automatic Control,2013,58(12):3112-3125.
[7]董文会.多摄像机监控网络中的目标连续跟踪方法研究[D].济南:山东大学,2015.DONG W H.Research of target continuous tracking inmulti-camera surveillan networ[D].Jinan:Shandong University,2015.
[8]谭飞刚,黄玲,刘建.多线索融合的无视野重叠跨摄像机行人跟踪算法[J].计算机应用研究,2017,34(8):2542-2545.TAN F G,HUANG L,LIU J.Pedestrian tracking across nonoverlapping camera views based on fusion multi-clue[J].Application Research of Computers,2017,34(8):2542-2545.
[9]蒋建国,顾占冰,胡珍珍,等.多摄像机视域内的目标活动分析[J].电子学报,2014,42(2):306-311.JIANG J G,GUZ B,HU ZH ZH,et al.Activity analysis cross muti-camera[J].Acta Electronica Sinica,2014,42(2):306-311.
[10]MEYER F,BRACA P,WILLETT P,et al.A scalable algorithm for tracking an unknown number of targets using multiple sensors[J].IEEE Transactions on Signal Processing,2017,65(13):3478-3493.
[11]SUNG Y,CHUNG W.Hierarchical sample-based joint probabilistic data association filter for following human legs using a mobile robot in a cluttered environment[J].IEEE Transactions on Human-Machine Systems,2016,46(3):340-349.
[12]HAN J,TAN T,CHEN L,et al.Moving objects representation for object based surveillance video retrieval system[J].International Journal of Security&Its Applications,2014,8(2):315-322.
[13]BOUCHRIKA I,CARTER J N,NIXON M S.Towards automated visual surveillance using gait for identity recognition and tracking across multiple non-intersecting cameras[J].Multimedia Tools&Applications,2016,75(2):1201-1221.
[14]BRACA P,GOLDHAHN R,FERRI G,et al.Distributed information fusion in multistatic sensor networks for underwater surveillance[J].IEEE Sensors Journal,2016,16(11):4003-4014.
[15]SHANG L,ZHAO K,CAI Z,et al.An energy-efficient collaborative target tracking framework in distributed wireless sensor networks[J].International Journal of Distributed Sensor Networks,2014,2014(2):1-18.
[16]SHI K,CHEN H,LIN Y.Probabilistic coverage based sensor scheduling for target tracking sensor networks[J].Information Sciences,2015,292(C):95-110.