Technology Transfer

Video-processing-based system for Automatic Estimation of Pedestrian Flow
  • Year:2019
  • Project:DOIT Program
  • Category:Imaged assist technology
  • Technique Features:

    In recent years, the development trend of vehicle technologies in the world is mainly based on the development of ADAS and ADS. In addition to vehicle safety assistance systems, pedestrian flow estimation is also a key part of smart transportation. Passenger vehicles or autonomous bus directly or indirectly require related information such as people counting applications as a dispatch assistance. ARTC also uses image processing and deep learning as the basis to build people counting technology, which will help improve the public transportation efficiency of vehicle scheduling to reduce personnel costs. While ARTC is studying autonomous vehicles, peripheral hardware and software have also been developed simultaneously, making urban transportation and development become more intellectual.

  • Specification:

    Detect rate : ≧80%
    False rate : ≦10%
    Frame rate: 800ms
    Detect range: 4.5m * 2.0m

  • Applications:

    Pedestrian flow statistics and analysis.

  • Benefit:

    In general, the surveillance system can be applied for several people counting applications, such as pedestrian traffic in public transportation and tourists flow estimation on bridge. For previously, the people counting method includes manual statistics, rotary gate of mechanical equipment, infrared optoelectronic equipment. However, these traditional methods have some problems which contain labor costs increasing, inconvenient passing through of pedestrian, and inexact counting in high-flow people place.

  • Contact:

    rdservice@artc.org.tw