基于改进进化匈牙利的多目标跟踪算法研究
|
摘要
为解决智能车在城市环境下多个障碍物同时跟踪的难点问题,提出一种基于改进进化匈牙利的激光雷达多目标跟踪算法。首先利用广度优先搜索算法划分出多个最大连接子图;然后采用进化匈牙利算法对非平衡问题下的最大连接子图序列依次求解、关联;最后将卡尔曼滤波跟踪过程实时优化,对关联上的障碍物进行平滑跟踪,对未关联的已有的障碍物进行预测,将未关联的新出现的障碍物添加到跟踪列表。实验结果表明:该方法可以很好地解决误关联问题,较多目标假设跟踪算法跟踪距离提升27%,跟踪精度提升35%,同时将卡尔曼滤波跟踪过程优化以后,速度准确率较传统卡尔曼滤波提升10%,检测精度显著提高。
To overcome the difficulty of intelligent vehicles in simultaneously tracking multiple obstacles in urban environment, the paper evolves the improved evolutionary Hungarian algorithm into a new multi-target tracking method. It utilizes the breadth first search algorithm to mark off several maximally-connected subgraphs, which are solved and associated under non-equilibrium state. After real-time optimization of Kalman filter tracking process, this method smoothly tracks the matched obstacles, predicts unmatched old ones, and adds unmatched new ones to the tracking list. Experimental results show that this method can effectively prevent matching errors, and increase the tracking distance by 27 percent and precision by 35 percent than multiple false targets tracking algorithm. Compared with the traditional Kalman filter algorithm, it has an obviously good detection capability due to an increase in speed accuracy by 10 percent after optimization of the tracking process.
To overcome the difficulty of intelligent vehicles in simultaneously tracking multiple obstacles in urban environment, the paper evolves the improved evolutionary Hungarian algorithm into a new multi-target tracking method. It utilizes the breadth first search algorithm to mark off several maximally-connected subgraphs, which are solved and associated under non-equilibrium state. After real-time optimization of Kalman filter tracking process, this method smoothly tracks the matched obstacles, predicts unmatched old ones, and adds unmatched new ones to the tracking list. Experimental results show that this method can effectively prevent matching errors, and increase the tracking distance by 27 percent and precision by 35 percent than multiple false targets tracking algorithm. Compared with the traditional Kalman filter algorithm, it has an obviously good detection capability due to an increase in speed accuracy by 10 percent after optimization of the tracking process.
引文
[1] 邹斌,刘康,王科未,等.基于三维激光雷达的动态障碍物检测和追踪方法[J].汽车技术,2017(8):19-25.
[2] 韩栋斌,徐友春,李华,等.基于手眼模型的三维激光雷达外参数标定[J].光电工程,2017,44(8):798-804.
[3] FERGUSON D,DARMS M,URMSON C,et al.Detection,prediction,and avoidance of dynamic obstacles in urban environments[C]//Intelligent Vehicles Symposium.Eindhoven.Netherlands:IEEE,2008:1149-1154.
[4] URMSON C,ANHALT J,BAGNELL D,et al.Autonomous driving in urban environments:Boss and the urban challenge[J].Journal of Field Robotics,2008,25(8):425-466.
[5] MERTZ C ,NAVARROSERMENT L E,MACLACHLAN R,et al.Moving object detection with laser scanners,JFR[J].Journal of Field Robotics,2012,30(1):17-43.
[6] ANNA PETROVSKAYA S T.Model Based Vehicle Detection and Tracking for Autonomous Urban Driving[M].[S.l.]:Kluwer Academic Publishers,2009,26(2-3):123-139.
[7] SVENSSON L ,SVENSSON D,GUERRIERO M,et al.Set JPDA Filter for Multitarget Tracking[J].IEEE Transactions on Signal Processing,2011,59(10):4677-4691.
[8] 杨飞,朱株,龚小谨,等.基于三维激光雷达的动态障碍实时检测与跟踪[J].浙江大学学报(工学版),2012,46(9):1565-1571.
[9] MA C ,HUANG J B,YANG X,et al.Hierarchical Convolutional Features for Visual Tracking[C]// 2015 IEEE International Conference on Computer Vision (ICCV).Santiogo:IEEE Computer Society,2015.
[10] PAN C,ZHANG J,QIN W .Real-time OHT Dispatching Mechanism for the Interbay Automated Material Handling System with Shortcuts and Bypasses[J].Chinese Journal of Mechanical Engineering,2017,30(3):663-675.
[11] 史璐璐,张索非,吴晓富.基于Tiny Darknet全卷积孪生网络的目标跟踪[J].南京邮电大学学报(自然科学版),2018,38(4):93-99.
[12] 柳毅,佟明安.匈牙利算法在多目标分配中的应用[J].火力与指挥控制,2002,10(6):34-37
[2] 韩栋斌,徐友春,李华,等.基于手眼模型的三维激光雷达外参数标定[J].光电工程,2017,44(8):798-804.
[3] FERGUSON D,DARMS M,URMSON C,et al.Detection,prediction,and avoidance of dynamic obstacles in urban environments[C]//Intelligent Vehicles Symposium.Eindhoven.Netherlands:IEEE,2008:1149-1154.
[4] URMSON C,ANHALT J,BAGNELL D,et al.Autonomous driving in urban environments:Boss and the urban challenge[J].Journal of Field Robotics,2008,25(8):425-466.
[5] MERTZ C ,NAVARROSERMENT L E,MACLACHLAN R,et al.Moving object detection with laser scanners,JFR[J].Journal of Field Robotics,2012,30(1):17-43.
[6] ANNA PETROVSKAYA S T.Model Based Vehicle Detection and Tracking for Autonomous Urban Driving[M].[S.l.]:Kluwer Academic Publishers,2009,26(2-3):123-139.
[7] SVENSSON L ,SVENSSON D,GUERRIERO M,et al.Set JPDA Filter for Multitarget Tracking[J].IEEE Transactions on Signal Processing,2011,59(10):4677-4691.
[8] 杨飞,朱株,龚小谨,等.基于三维激光雷达的动态障碍实时检测与跟踪[J].浙江大学学报(工学版),2012,46(9):1565-1571.
[9] MA C ,HUANG J B,YANG X,et al.Hierarchical Convolutional Features for Visual Tracking[C]// 2015 IEEE International Conference on Computer Vision (ICCV).Santiogo:IEEE Computer Society,2015.
[10] PAN C,ZHANG J,QIN W .Real-time OHT Dispatching Mechanism for the Interbay Automated Material Handling System with Shortcuts and Bypasses[J].Chinese Journal of Mechanical Engineering,2017,30(3):663-675.
[11] 史璐璐,张索非,吴晓富.基于Tiny Darknet全卷积孪生网络的目标跟踪[J].南京邮电大学学报(自然科学版),2018,38(4):93-99.
[12] 柳毅,佟明安.匈牙利算法在多目标分配中的应用[J].火力与指挥控制,2002,10(6):34-37