Automated Helmet Violation Detection and Registration Capture

Due to the affordability of motorcycles and their use as a daily mode of transport, there has been a significant rise in motorcycle accidents. This increase is largely attributed to the fact that many motorcyclists do not wear helmets, creating a constant danger when traveling by motorcycle. In recent years, the government has made it a punishable offense to ride a motorcycle without a helmet. Although the current video surveillance system is effective, it relies heavily on human assistance, which tends to diminish over time, and human bias can also affect its efficiency. Therefore, automating this process is highly desirable. In this paper, we present a method for the automatic detection of motorcyclists without helmets using real-time surveillance videos. The proposed method first identifies motorcycles in the surveillance footage through background subtraction. It then differentiates between helmeted and non-helmeted riders using first and second order derivative edge detection algorithms along with a neural network. If motorcyclists are detected without helmets, the system will capture the vehicle’s number plate using Optical Character Recognition (OCR) and a neural network. The project not only identifies the bike’s number plate but also digitally retrieves the plate number and saves the frame where the rider is seen without a helmet, which can serve as evidence against the biker. This enables the government to issue fines to the rider and assures society that there is oversight on the roads, even when they appear empty.