Open AccessReal‐time detecting and tracking of traffic shockwaves based on weighted consensus information fusion in distributed video network
Author(s) -
Yang Deliang,
Chen Yangzhou,
Xin Le,
Zhang Yuan
Publication year - 2014
Publication title -
iet intelligent transport systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.579
H-Index - 45
eISSN - 1751-9578
pISSN - 1751-956X
DOI - 10.1049/iet-its.2013.0038
Subject(s) - computer science , tracking (education) , sensor fusion , information fusion , fusion , artificial intelligence , real time computing , distributed computing , computer network , computer vision , psychology , pedagogy , linguistics , philosophy
Tracking traffic shockwaves (queuing shockwave and discharge shockwave) in a road section between intersections can be applied to obtain various traffic parameters, such as the queue length, stop delay etc. In video processing of a distributed low‐angle video network installed above the road section, various factors affect the accuracies of positioning of vehicles and tracking of traffic shockwaves. To overcome these effects, they propose a method of weighted consensus information fusion to track the traffic shockwaves in real time. In the visible region between opposite cameras, the cameras detect the shockwaves through the duplex flexible window fused with AdaBoost cascade classifiers and meanwhile dynamically estimate the weight of the measurement noise. In the blind region between contrary cameras, the cameras use the speed changes of the vehicles entering and leaving the blind region to estimate the shockwaves’ positions. Thus, by exchanging the information among the cameras through communication and dynamically adjusting the confidence level of the detected results, the algorithm of weighted consensus information fusion effectively obtains globally optimal estimation of the shockwaves. Experimental results show finer tracking results of the shockwaves during morning and evening rush hours by the proposed method.