Seatbelt Automatic Recognition Algorithm Based On Image Analysis

Intelligent Transportation System(ITS)is a research topic in the world.In this paper,image processing and pattern recognition is applied to the intelligent transportation field,which is of great significance to the increasingly serious traffic problems.At present,with the increase of road bayonet camera,automatic license plate recognition,red light running automatically capture recognition speeding recognition have been automated,and not wearing a safety belt automatic identification has just started,at present is mainly by artificial screening,the workload is very large.In recent years,due to the number of traffic accidents is increasing,attracted the attention of more and more of the related department,safety belt,as an important means of guarantee the driver and passenger safety and lower the mortality rate,the traffic control department strict requirements of drivers and passengers in the vehicle driving process to wear seat belts,and this is also our country traffic law clearly stipulates the.In view of the above situation,this paper designs a set of automatic identification system based on image analysis.After a lot of experiments,the effect is very satisfactory.If the real input is used,it will lighten the burden on the staff.The whole system is divided into 4 parts:(1)the image of the acquisition to the first preprocessing,such as gray,binarization(in this paper,a new image of binarization method).(2)the location of the license plate,the texture features of the license plate,combined with morphological operation method,the character of the mass detection,and then to the character,direction of the clustering,so as to realize the license plate location.(3)the upper and lower boundary of the window glass is calculated by the position and the coordinates of the vehicle plate,and the horizontal and vertical directions are calculated.The driver is located by the color of skin color and the position of the steering wheel.(4)using the Canny operator,the edge of the image,and then use the probability of Hof transform and least square method to fit a straight line,and extract the feature vector,and finally the support vector machine to classify,determine whether the driver wearing a seat belt.This experiment proves that the method is feasible and effective in time and accuracy.