Automated Accident Detector Using Deep Learning

CCTV surveillance has evolved to be an important tool in detecting and following vehicles, both individual and public transport. This advantage has highly impacted in various application such as safety of drivers, commoners by tracking any crime or monitoring the accidents. Road accidents are alarmingly increasing in the busy streets of our country. Our project mainly focuses on automated detection of road accidents by CCTV and immediate information is passed to the concerned people. As a result, we can treat the injured people quickly and save many lives by instant action. To identify car accidents, our technology combines an algorithm for deep learning with an object recognition and tracking mechanism. Here, object detection is done using YOLO. This can track and monitor the vehicles that exhibit unexpected events that primely cause collision between other vehicles. The technique called YOLOv7 is also used to perform the same to compare the accuracies. This can integrate object categorization and bounding box prediction into a single entirety process.