Helmet Detection and Number Plate Extraction

Author Name : Het Trivedi, Dr. Ankita Karale, Dhir Machhi, Vivek Patel, Aniket Jagtap

ABSTRACT

Existing systems monitor traffic violations primarily through CCTV recordings. This requires the traffic police to examine the frame in which his traffic violation occurs and zoom in on the number plate if the driver is not wearing his helmet. However, with traffic violations occurring frequently and people using motorbikes increasing daily, this requires a lot of manpower and time. What if there was a system that automatically checks for traffic violations when he is riding a motorcycle/moped without wearing a helmet and if so automatically he extracts the license plate number of the vehicle. Recent work has successfully done this work based on CNN, R-CNN, LBP, HoG, Yolo function and others. However, this work is limited in terms of the efficiency, accuracy, or speed with which object detection and classification are performed. In this study, we build a non-helmet driver detection system that attempts to achieve automation of detecting traffic violations of not wearing a helmet and extracting vehicle license plates. Using the methods described above, we constructed a comprehensive system for both helmet detection and license plate extraction.

Keywords— Image Processing, Classification, Conversational Neural Network, YOLOv2, YOLOv3, OCR, Motorcycle Riders, Deep Learning