One of the most concerning factors arising out of traffic overflow is the increase in number of accidents due to violations of traffic rules leading to severe losses with life money and physical disabilities. With objective of prevention of accidents, the smart image capturing devices monitoring traffic internal within vehicles and CCTVs on road premises of present days requires algorithms and hardware modules that detect the vehicles violating traffic rule and avoiding possible mishaps due to negligence through instant alerts. This research proposes a comprehensive system utilizing YOLO (You Only Look Once) object detection algorithm integrated with NodeMCU microcontroller. This NodeMCU is interfaced to motor driver of vehicle with display and Internet of Things (IoT) capabilities. Motor driver controls the motor output of the vehicle. YOLO is selected for its high accuracy and fast inference capabilities. This enables identification of real time road hazards from vehicles, pedestrian’s cyclist and traffic signs captured by mounted cameras in vehicles. The performance of the system is evaluated through accuracy, frame rate, reaction time, motor response efficiency.
Object Detection using Machine Learning to Avoid Accident
2024-11-18
693723 byte
Conference paper
Electronic Resource
English
Traffic Accident Detection Using Machine Learning Algorithms
Springer Verlag | 2022
|Accident Investigation: Training to Avoid Rushing the Judgement
British Library Conference Proceedings | 1994
|Accident Detection Using Deep Learning
IEEE | 2020
|