Accident Detection and Alert Generation


Vedika Sadavarte, Mrs. Priyanka More,
Department of Computer Engineering, Vishwakarma Institute of Information, Technology, Survey No. 3⁄4, Kondhwa Budruk, Pune, Maharashtra, India.

Ramsha Shaikh,
Department of Information Technology, Sandip Institute of Technology and Research Centre, Nashik, Maharashtra, India.

Dr. Sachine Sakhare, Mrs Snehal Rathi,
Department of Computer Engineering, Vishwakarma Institute of Information Technology, Survey No. ¾, Kondhwa Budruk, Pune, Maharashtra, India.


The current scenario of increasing vehicles has resulted in traffic hazards and road accidents. The continuous growth of road accidents has also caused an increase in casualties due to accidents. One of the primary reasons for the increased rate of casualties is due to lack of emergency services. The delay happens due to traffic congestion and unstable communication with medical units. The proposed system aspires to deliver a self-operating accident detection system with alert generation to provide timely aid in crucial ways. A surveillance system enriched with the concept of deep learning technique of Convolution Neural Network built to detect real-time accidents and future communicate the details by generating an alert on a web application which will be handled by medical units and police authorities. The system also delivers the functionality to report accidents manually in remote areas. To train and evaluate the detection system we collected 120 video frames under numerous conditions. The experimental results exhibit that the proposed system can detect an accident and send an alert on the web app with a mean absolute percentage error that is less than 20%.