An Effective Addressal of Traffic using Tensor Flow

Effective city surveillance using CCTV with automated accident alarm to the adjacent police station and automatic death alert based on the nearby police station Keep a record or log to indicate if action has been done and the specifics of that activity. Whether the accident be more severe or mild there's somehow need for a medical assistance. So, the system automatically arranges a call out for ambulance to the location. Instead of using a single model for detection of accident, severity and vehicle type we are going to use multiple models which follows a hierarchy in this way the computation power needed for prediction will be optimized compared to a single model. Themodel that are built and trained are CNN which takes images as an input if anaccident is detected than the image is fed to the next two models or else it will predict for the next image. If accident is detected, then the image is fed to the next model to know the accident severity and vehicle type if severity is medium or high than the algorithm sends the details to the police station and calls an ambulance and record it in the database afterward it will turn on the recording mode (it will retrieve video 5 minute prior to the accident till the ambulance arrived) after the detection of the ambulance the Accident prediction cycle continuous.