China Safety Science Journal ›› 2023, Vol. 33 ›› Issue (S1): 138-144.doi: 10.16265/j.cnki.issn1003-3033.2023.S1.0132

• Safety engineering technology • Previous Articles     Next Articles

Research on 3D real scene reconstruction for road traffic accidents

ZHENG Jinzi1(), YANG Qi2, LIU Jun1, LI Liang3, CHAI Yanan1, XU Pengfei1   

  1. 1 Research Institute for Road Safety of MPS, Beijing 100062, China
    2 Detachment Five of Traffic Police of Shanxi Public Security Expressway, Xinzhou Shanxi 034000, China
    3 Wuhan Daspatial Technology Co., Ltd., Wuhan Hubei 430000, China
  • Received:2023-03-02 Revised:2023-05-10 Online:2023-06-30 Published:2023-12-31

Abstract:

In response to the shortcomings of traditional methods for scene investigation of road traffic accidents, such as long time consumption, difficulty in re-exploration, and low accuracy, the technology of three-dimensional (3D) real scene reconstruction for road traffic accidents was explored. Combined with the application experience of 3D real scene reconstruction in relevant fields, the following research work including field image acquisition and internal model construction was carried out. In response to the diverse geological forms, uncertain scale, and short duration of traces at road traffic accident scenes, a series of UAV tilt photography was proposed for different types of accident scenes. The algorithm of 3D real scene reconstruction for traffic accident scene modeling was designed, and the precision inspection method of accident scene survey based on 3D real scene model was developed. A single vehicle rollover accident on Shanxi highway was selected as a case study to construct a 3D real scene model of this accident and to verify its accuracy. Research has shown that the survey method for road traffic accidents based on a 3D real scene model can reduce the survey time by more than half and facilitate re-survey after traffic recovery. Besides, the constructed 3D model of the accident scene has an accuracy of centimeter level, and the measurement error meets the industry standard of Investigation of Road Traffic Accident Traces and Evidence issued by the Ministry of Public Security.

Key words: road traffic, accident scene investigation, 3D real scene reconstruction, tilt photography, unmanned aerial vehicle