Sentosa Passenger Counting System

Enhancing the efficiency and safety of public transportation is crucial, and automatic passenger counting (APC) and seat occupancy detection play a significant role in achieving this goal. However, current APC methods, which rely on infrared and pressure sensors or vision, face challenges in accuracy, installation, and occlusion handling, particularly for low-top, open-air sightseeing trams. Although some studies have explored the use of 3D-optical time-of-flight (TOF) sensors and millimeter-wave radar, the application of 3D LiDAR in APC remains largely unexplored. To address these challenges, we present a novel algorithm that integrates 3D LiDAR and RGB cameras to accurately count passengers and detect seat occupancy in open-air sightseeing trams. Our proposed system holds significant promise for enhancing the management and operation of sightseeing trams in tourist areas, elevating visitor experiences, and fostering sustainable tourism. 2023 IEEE International Conference on Cybernetics and Intelligent Systems (CIS) and IEEE Conference on Robotics, Automation and Mechatronics (RAM)2023APC: Automatic Passenger Counting and Seat Occupancy Detection for Sightseeing Tram based on 3D LiDAR and CameraYiyao Liu, Qiyang Lyu, Jun Zhang, Xudong Suo, Chaoyu Hong, Lin Lin, Danwei WangDOI

Highlights

• Superior bandwidth control: Superior bandwidth control: 5G-based, low-latency image transmission from the tram to the MEC, with 11 depth cameras running at 25 FPS requiring less than 1 Mbps.
• Complete monitoring solution: An end-to-end framework that allows users to access all information directly through a GUI.
• Robust and accurate counting: Day-and-night, all-weather tram detection with 99.9% precision and 95.9% recall.
• Easy deployment: Hardware installation and tram retrofitting can be completed in less than one day.

System framework

Demo