A Sensor Fusion and Yolo-Based Approach for Real-Time Vehicle Accident Detection

Abstract

Road accidents remain a critical global issue, often resulting in severe injuries and delayed emergency assistance, particularly in remote and high-speed environments. This paper proposes a smart accident detection and monitoring system that combines Internet of Things (IoT) devices with artificial intelligence techniques to enhance detection accuracy and response time. The system employs an ESP32 microcontroller integrated with an MPU6050 accelerometer to continuously monitor vehicle motion and identify sudden impacts based on predefined thresholds. To improve reliability, GPS-based speed analysis is incorporated to validate critical events. In addition, an OV2640 camera module captures real-time images, which are analysed using the YOLOv8n deep learning model for visual verification of accident conditions such as vehicle damage or fire. Upon confirmation, the system retrieves precise location data using a GPS module and transmits alert messages containing map coordinates to emergency contacts via a GSM module. Experimental results indicate improved detection accuracy and reduced false alarm rates compared to conventional single-sensor approaches. The proposed system is cost-effective, scalable, and suitable for real-time deployment, contributing to faster emergency response and enhanced road safety.