Search

Object must be selected manually and scale variation can't be handled with properly when using traditional algorithms. To overcome this problem, this paper proposed a new object tracking algorithm fused with optical flow detection and template matching. Firstly, it automatically detected and extracted object by combining the optical flow and the image segmentation results, accomplishing detection based tracking. It used the template matching technique for target locating when the result of the detection based tracking was unreliable, accomplishing matching based tracking. Finally, with the help of the template's auto update, adapted the target's changing in scale by the tracking window. The experimental results demonstrate that the proposed method can not only adapt object's scale variation, but acquires stable tracking results. Compared with three other algorithms, the proposed method has advantages in automatic detection and scale self-adaptation. [ABSTRACT FROM AUTHOR]

According to the coordinated control problem of urban region transportation, this paper presented a traffic cloud computing platform, in which considered the relationship between different control nodes, to allocate control strategy for getting more reasonable the signal timing for different intersections in urban transportation, the growing Internet of things provided reference recommendations on the development strategy for transportation managements based on the characteristics of dynamic change of execution status among different control nodes and the random mutations in running of cloud computing platform. The experimental result shows that the proposed coorderate control cloud strategy is not only increase the traffic efficiency, but also reduce significantly the running energy consumption of cloud computing platform. [ABSTRACT FROM AUTHOR]