Document Type : Original Article
Faculty of Computer Engineering and IT, Shahrood University of Technology, Shahrood, Iran
Computer Science Department, Faculty of Mathematics and Computer, Higher Education Complex of Bam, Bam, Iran
Helmet are essential for preventing head injuries in bikers. Traffic laws are applied in most countries to bikers who don’t wear a helmet. Manually checking bikers for the usage of a helmet is a very costly and tedious task. In this regard, several helmet detection methods were developed in literature for detecting bikers violating the law in recent years. This paper proposes an image processing method based on the Local Binary Pattern (LBP), Local Variance (LV), and Histogram of Oriented Gradient (HOG) descriptors for detection of bikers without a helmet. The innovation of the proposed method is mainly on the feature extraction step, which leads the classification towards appropriately discriminating between the two classes of helmet and non-helmet. The experimental results show our method is superior to the existing methods for helmet detection. The accuracy of the proposed helmet detection method is 98.03% using the Support Vector Machine classifier.