Object Detection for Video Surveillance Using Edge-Cloud Collaboration

Road safety and emergency response are greatly improved by accurate and prompt accident detection. This study describes a novel method for accident detection that makes use of a Telegram bot for communication and a Raspberry Pi as an edge device. Based on video feeds, the system uses computer vision techniques, notably the YOLOv3 algorithm, to detect accidents in real-time. The system instantly notifies chosen receivers, such as emergency services or worried parties, via the Telegram bot when an accident is detected. Results from experiments show how well the system performs in terms of reliably recognizing accidents and sending out timely alerts. The system's real-time execution, useful applications, and highlighted restrictions offer insightful information for upcoming improvements. This paper offers a scalable and flexible solution for enhancing emergency response and road safety, which advances the field of accident detection and management.