Link to Content Area
:::

Institute of Transportation, MOTC

:::
  • small size
  • medium size
  • large size
  • print
  • facebook
  • plurk
  • twitter

Application Scenarios and Pilot Run of Vehicle and Pedestrians Traffic Flow Information Using UAV Aerial Videography (2/3) - Non-signalized Intersections

  • Date:2023-09-20
  • Update:2023-09-20
  • Department:IOT

Project Overview
       Previously, improving traffic safety meant analyzing accident data after the fact. In recent years, the Institute of Transportation has developed traffic conflict analysis tools that, when combined with UAV aerial photography and artificial intelligence (AI) image recognition technology, automatically convert pedestrians at intersections and vehicle flow images into trajectory data. In order to
prevent accidents from occurring, software is used to analyze the venues and patterns of intersections prone to traffic accidents. This is the second year of a three-year plan between 2021 and 2023 to apply and validate analysis tools. In addition to the two pilot test projects, four accident-prone intersections were chosen for analysis and will serve as a reference for local governments to improve
intersections in the future.
Research Results

  1. In terms of object detection technology, aerial videos were compressed to reduce processing time and conduct research on pedestrian detection, bicycle classification status, and vehicle detection frame deflection issues.
  2.  Upgrades to the software optimization trajectory/vehicle speed drawing function, conflict hot zone image supposition, and conflict video playback function have all been analyzed.
  3.  13 roads/intersections have been chosen in collaboration with the Directorate General of Highways, MOTC, New Taipei City, and Taichung City to carry out pilot test projects, namely "right turn conflict" and "non-signalized road/ intersection stop and yield conflict." Relevant functions have been added to the analysis software in addition to proposing risk driver behaviors and determining logic.
  4.  Four accident-prone roads/intersections were chosen for analysis by Keelung City, Hsinchu County, Miaoli County, and Changhua County. Local governments have been given analysis data to use as a reference for future road/ intersection improvements.

Result Promotion and Benefits

  • Two pilot test projects have been completed: "right turn conflict" and "non-signalized road/intersection stop and yield conflict." The analysis data of 13 roads/intersections was provided to the Directorate General of Highways, MOTC, New Taipei City, and Taichung City for reference and subsequent improvement.
  • A traffic conflict analysis of four accident-prone roads/intersections has been completed. The analysis data has been provided as a reference for future road/intersection improvements in Keelung City, Hsinchu County, and Changhua County.
  • On October 27, 2022, the proposal's result presentation and education training seminar were held to advocate for research results.

Summary of Research Results

47

Analysis of the right-turn conflict situation at roads/intersections using right-turn traffic flow trajectory

Research Result Report
Application of Road Traffic Flow Analysis and Accident Detection  (Published in November 2019)
•  Development of Traffic Collision Analysis Technology using UAV Aerial Videography at Intersections  (Published in July 2020)
• The Case Application and Comparison of Road/ intersection Overhead Photography in Traffic Conflict Analysis (Published in July 2021)
Application Scenarios and Pilot Run of Vehicle and Pedestrians Traffic Flow Information Using UAV Aerial Videography (1/3) – Establishing a Test Plan (Published in May 2022).
• Application Scenarios and Pilot Run of Vehicle and Pedestrian Traffic Flow Information Using UAV Aerial Videography (2/3) – Unsignalized Intersections  (Published in May 2023).

Count Views:80
Top