In present situation, deaths due to accidents are increasing very fast. The full picture for the happened accidents are drunken driving, rash driving, carelessness,drowsiness, etc. Most of the accident demises occur due to not proving instant medicinal care in time to the accident site and carelessness. Anyone is not taking any action when they saw an accident. The main intent this work is developing a system that will detect the accident using CCTV footage. From the stream of video images are extracted. To classify images Convolutional Neural Networks (CNN) incorporated. This system detects the accident if occur. Since CNN proven to be a fast and precise method to classify images, it detects the accident with more accuracy. For smaller datasets it gives more than 95%. It also requires less processing time when compared to other image classifying algorithms.

Deep Learning, Convolutional neural networks, Accident detection, CCTV footage.