Traffic Sign Detection And Recognition Under Haze Weather

Coupled with the rapid development of intelligent transport systems,road traffic sign recognition as an important part of intelligent transportation system has become the current research focus.However,in north and northeast areas of China where often haunted the bad fog weather that make the existing traffic sign detection system’s recognition and accuracy rate greatly reduced,which unable to meet the demand of reality.In order to improve the environment of haze’s traffic sign’s recognition rate,this paper designed a traffic sign recognition system.Firstly,detecting haze is the first part of a sign recognition system in the haze weather.By researching differences between the clear image and haze image in the same scene,we can get these four characteristics of haze image: image brightness,image contrast,image recognition and oneness of haze areas.Thus,Based on the image brightness and image contrast,a Fisher classifier is designed that aimed at judging whether the current image is a haze image.Then,once an image is considered as haze image,the image deblurring needs to be executed.The research on enhancement of haze images mainly includes two aspects: one is based on mathematical models,another based on human visual characteristics.This thesis outlines some classic algorithms of the above two aspects.The simulation of these algorithms is accomplished on MATLAB.Then we summarize results of the simulation.In this paper,the image restoration technology is used to build the model of the atmosphere light attenuation and ambient light.Their interactions generate the imaging model in a haze weather.Thus,a clear image can be obtained via reverse derive.Finally,the traffic sign detection and recognition algorithm is researched which includes sign detection and sign identification.Based on color information segmentation and shape feature location’s traffic sign detection and recognition,we propose an improved algorithm.Typically in the sign identification,adopted template matching algorithm which shortens the time is performed to identify the traffic signs.This paper designs a traffic sign detection and recognition system under the haze weather.It can greatly improve detection accuracy in the haze environment.Besides,this system is also applied in the sunny environment.