Deep Learning-based Auto Accident Detection and Alert System for Vehicles

The increasing number of vehicles, especially in urban areas, has led to a rise in traffic violations and accidents. In India, accidents are a leading cause of death, with over 75% of fatalities resulting from the inability to provide prompt assistance. To address this issue, the objective is to develop a system that can identify accidents using CCTV footage and send alerts to the nearest control room. The system utilizes deep learning algorithms such as VGG16 for image classify video frames as accidents or no accidents, these neural networks have shown to be quick and precise at image classification, showing a 95% higher accuracy rate than other algorithms, while also requiring less pre-processing.