Published Paper Details:

The rate of increase in the number of two-wheelers in India is 20 times faster than the rate of increase in the human population. Riders who do not wear a helmet have a 2.5-fold increased risk of fatality compared to those who do. The current video surveillance-based system is effective, but it requires a lot of human intervention, which reduces its efficiency over time and introduces human bias. This project tries to address this issue by automating the process of detecting motorcyclists who are not wearing helmets while riding. The system detects moving objects in the scene using a video of traffic on public highways as input. To determine whether the moving object is a two-wheeler, a machine learning classifier is applied to it. If the rider is on a two-wheeler, a different classifier is employed to determine whether or not the rider is wearing a helmet. As a result, wearing protective headgear is crucial to reducing the risk of injury in the event of an accident. This project offers a mechanism for locating individual or several riders who are riding motorcycles without helmets. Bike riders are identified with the use of the YOLOv3 model, which is a consistent type of YOLO model, from the starting stage of the proposed approach. The frontline methodology for object distinguishing aids as such in identifying the riders with and without helmet. If the number of riders reaches two, the vertical projection of a binary picture is utilized to count them. Along with this speed and drowsiness added. A message module is sent to over speed and for drowsiness alarm is produced in inactive mode.