Real-time Sensing and Surveillance


NICTA is developing advanced sensing and surveillance technologies capable of providing rich information about traffic states to a new breed of smarter traffic control applications.

  • Accurate traffic flow measurement and vehicle classification together with stopped vehicle detection and the identification of traffic incidents and anomalies will offer a finer level of control to adaptive traffic control systems and provide for more rapid response to traffic incidents.

  • Extensive real-time test beds established on the streets of Sydney support the analysis and semantic level understanding of traffic at intersections and on motorways. In seeking solutions which afford low-cost installation and maintenance, the project is developing solutions which can make effective use of constrained position/angle cameras mounted on existing traffic signals under a range of weather and lighting conditions and the presence of heavy shadow, occlusion and congestion.  

  • Reliable, low cost vehicle tracking, queue length estimation and vehicle classification at metropolitan and urban intersections and on motorways will provide valuable additional information enabling advanced traffic management and control systems to ease congestion, reduce variability in travel times and improve intersection safety.

  • Truly intelligent vision systems for traffic monitoring, remove the burden on human operatorís to detect and analyse traffic incidents and increas etheir capacity for higher-value tasks such as responding to the event itself.
The major areas of research include object feature extraction, invariant pattern recognition and multi-sensor data fusion in order to obtain accurate, real-time information about individual vehicles and groups of vehicles and traffic states to better inform real-time traffic control decisions.

NICTA sensors will  be mounted at traffic signal height for ease of installation and maintenance and be effective in a range of weather and lighting conditions.  Using data fusion techniques and enjoying a close coupling with the traffic modelling and control systems, traffic analysis will be impervious to the impact of vehicle occlusion which might otherwise diminsh the effectiveness of street-level video sensing.

Best Paper
Vehicle classification at nighttime using Eigenspaces and Support Vector Machine PDF_icon
Tuan Hue Thi, Kostia Robert, Sijun Lu and Jian Zhang, in Proc. IEEE International Congress on Image and Signal Processing, 2008, Sanya, China

STaRSense Night


  • Journal and International conference publications


  • 6 patents and patent applications


  • 8 active PhD students
  • 20 active under/post grad students


  • Organiser IEEE MMSP08 Workshop
  • 10 PC committee members and workshop chairs @ MMSP08

RTA_logoNICTA's Smart Transport and Roads Project is a collaboration with the Roads and Traffic Authority of New South Wales, developer, supplier and end-user of the world leading SCATS traffic management system.



Dr. Glenn Geers,
   Project Leader,
   +612 8306 0439