| Data shows that traffic safety is one of the rough problems in today’s world traffic field. In order to reduce transport accidents caused by human, improve road management level, more and more intelligent transportation technologies are applied into practice. Detection and recognition of road traffic sign based on the video image are an important part of intelligent transportation system. Common road traffic signs mainly refers to zebra crossing, turn markers and lane which provide lane, turn and stop information for intelligent vehicles so as to provide protection for the safety driving of intelligent vehicle.This paper mainly focuses on the characteristics of the road image researching in the extraction and recognition of the traffic signs in the road image. The main work includes the following aspects:Firstly, perform the pretreatment for the road images from videos, which includes image reduction, gray, light adjustment, Ostu binary.Secondly, research on the detection of the common road traffic signs.Specifically, we extract the interest region of the road image and then detect traffic signs according to the characteristics of different traffic signs.Finally, identify the turn of the turn marker detected and do further processing for the lane detected, which include line fitting for the lane near the zebra crossing and connection for virtual lane, etc.In this paper, some prior knowledge of road traffic signs is applied into the detection algorithm that can not only improve the efficiency of the algorithm of detection and recognition but also has good execution efficiency. |
