Driving Assistance : Pedestrians and Bicycles Accident Risk Estimation using Onboard Front Camera

In this study, we propose a collision detection system by detecting and estimating the risk posed by pedestrians and bicycles in the images captured by a monocular onboard camera. In the proposed intrusion system. after initial detection, the pedestrians and bicycles are tracked to obtain their location, the direction of movement, and posture information using lane detection information, velocity calculation and, pose estimation respectively. Finally, this information is evaluated using fuzzy rules to estimate the risk the pedestrian and/or bicycle poses. The results are transmitted to the driver using voice and sound. We tested the system using 89 video scenes and achieved recall and precision accuracies of 0.94 and 0.87 respectively.

This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Natureʼs AM terms of use (https://www.springernature.com/gp/open-research/policies/accepted-manuscript-terms), but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s13177-022-00324-2