Helmet detection on two-wheelers and number plate recognition using image processing

Abstract

There has been a sudden increase in motorcycle accidents over the years. The helmet is a safety equipment that protects the motorcyclists, however many riders (i.e. students) don’t use it and the results could be fatal. This paper holds an agenda to propose a system for detection of motorcyclists without helmet. The proposed idea first detects motorcycle riders using surveillance video using background subtraction and object segmentation methods. We have also used techniques that involve human face detection using image processing. This includes face detection using Haar like feature technique and circular Hough transform method that helps in detecting any circular object hence detecting the helmet. This approach works in the near real time mode with significantly low false alarms.