Provision of High Accuracy Vehicle Classification and Counting System in Government Tunnels

I&T Wish Provision of High Accuracy Vehicle Classification and Counting System in Government Tunnels
(REF : W-0413)
Matched I&T Solution
Trial Project
Summary and Challenges
  • Due to toll waiver arrangement in Lantau Link and Tseung Kwan O tunnel, the traditional vehicle counting by toll will not be available and it raises an issue to provide the vehicle counting function by other alternatives. Furthermore, the existing data gathering and submission process by toll involves too many manual works for data processing. It leads to human mistakes easily and a consumable processing time. From the news reports in the past years, there were cases of unauthorized access of pedestrians or bicycles which cause danger and potential traffic accident.
  • In view of the aforementioned problems, it comes to the needs for a new and complete solution of High Accuracy Vehicle Classification and Counting in Government Tunnels with the following features:
    • To provide a high accuracy objects classification for counting vehicle and detecting invalid objects in order to enhance the tunnel monitoring.
    • To make use of video camera and LiDAR to optimize the class detection classes.
    • The accounting of vehicle flows shall be flexible and able to adopt to tunnel operation such as temporary traffic arrangement.
    • The vehicle analytics data shall be automatically uploaded to Government’s application servers by means of GWIN.
Expected Outcome
  1. To develop an AI model and integrate with video cameras and LiDAR for tunnel traffic analytics, and achieve a High Accuracy Vehicle Classification and Counting with over 98% in accuracy.
  2. To provide a real-time traffic monitoring and a comprehensive Vehicle Detection Classes including Motorcycle, private car, taxi, van, light bus, truck, bus, etc.
  3. To identify invalid objects such as unauthorized pedestrian or bicycle in order to enhance safety and notify the tunnel operation staffs.
  4. To provide estimation of journey time, travelling speed and detection of cross-lane event by tracking the vehicle sequence inside tunnel.
  5. To provide a fully 7/24 automatic data processing and generate a comprehensive report for public transportation analysis.
Expected Trial Duration 9-month
Contact Information
I&T Wish Proposer:Electrical and Mechanical Services Department (EMSD)
Contact Person:LEUNG Ka Chun
Position:Engineer/Transport Services3/2
Tel: 3908 2565
Initiating Department Transport Department (TD)
Upload Date 2022-04-29
Closing Date 2022-05-13