Analytical Modeling to Estimate Vehicle Speed in Forensic Video for In-Depth Investigations of Traffic Accidents
Qiang Chen, Haijing Hou, Zhiwei Guan, Da Li · CICTP 2021 · 2021
Currently, estimation of vehicle speed in forensic videos is becoming increasingly important for analyzing traffic accidents. To improve the efficiency and scientific nature of this study, an analytical model is proposed. In this method, video code information is detected to calculate the time of video based on its frame rate and moving scale for calculating distance. Next, a distance interpolation method is established according to video status. Then, a velocity curve is created to analyze the braking deceleration of the vehicle. Finally, as a proof of concept, the proposed analytical model is implemented for simulation cases and actual traffic accidents. Experimental results demonstrate the feasibility of the method in terms of accurate estimation of vehicle speed. The results indicate an average relative error of less than 3%. This analytical model is highly versatile and can be used for in-depth investigations of traffic accidents.