Smart Technology and Artificial Intelligence
Cloud-Enabled Smart Surveillance System for Real-Time Car and Fire Accident Detection
Authors: Pogula Sreedevi, Shaik Habeeb, Sudamalla Subbalakshmi, Syed Faizan, K.Lakshmi Rupa and Karur Dadakhalandar
Publishing Date: 04-06-2026
ISBN: 978-81-975670-2-5
Abstract
Rapid detection of traffic collisions and fire accidents is essential for urban safety, yet traditional surveillance relies on expensive, difficult-to-scale edge devices. This study proposes a serverless computer vision system built exclusively on AWS to offer a scalable and cost-efficient alternative. The architecture utilizes Amazon S3 for video ingestion, AWS Lambda for frame extraction, and Amazon Rekognition— leveraging a ResNet-based Transfer Learning model—for real-time accident detection. To ensure reliability, the system was validated against a manually curated dataset of 292 images featuring challenging lighting, weather, and overhead camera perspectives. Experimental results demonstrate an overall Average Precision of 0.605, a Recall of 0.530, and an F1-Score of 0.563, with fire detection specifically achieving a Precision of 0.684. Upon detection, Amazon SNS successfully dispatches emergency notifications within 1.5 to 3 seconds. These findings confirm that an event-driven cloud pipeline is a viable solution for smart city surveillance, providing superior scalability and resilience compared to traditional edge computing models.
Keywords
Smart Surveillance, AWS Cloud, Accident Detection, Serverless Computing, Convolutional Neural Networks, Object Detection.
Cite as
Pogula Sreedevi, Shaik Habeeb, Sudamalla Subbalakshmi, Syed Faizan, K.Lakshmi Rupa and Karur Dadakhalandar, "Cloud-Enabled Smart Surveillance System for Real-Time Car and Fire Accident Detection", In: Mukesh Saraswat, Sandeep Kumar, Manjunatha Sughaturu Krishnappa and Rakesh Keshava (eds), Smart Technology and Artificial Intelligence, SCRS, India, 2026, pp. 12-20. https://doi.org/10.56155/978-81-975670-2-5-2