Abstract: With the increasing population, the number of two-wheeler riders has increased rapidly. This causes an increase in accidents rates. Riders, being careless by not wearing helmets might lead to more health damage. Human involvement in monitoring the vehicles may be less effective and hard to catch every rider with no helmet, thus monitoring all these manually may be difficult if we consider the huge number of riders. Observers may not be able to capture everyone with no helmet. Thus to ensure that every single rider is monitored properly a Computerized Automatic Helmet Detection System is implemented. The system is trained with a dataset for better accuracy and prediction of results. The system will detect the rider with no helmet and will extract the number plate of the two-wheeler. The system uses GUI and various Python libraries such as sklearn, NumPy, TensorFlow, Keras, etc, also algorithms such as Yolo V3(You Look Only Once) and SSD (Single Shot Detector) has been used for object detection.
Keywords: Helmet, Number plate, Detection, Python Libraries, Machine Learning, Yolo V3, SSD.
|
DOI:
10.17148/IJARCCE.2022.11361
