Design Of Highway Fog Early Warning System Based On Video Analysis Technology

With the continuous increase of the mileage of our country's expressways,the vehicles encounter the haze during the driving process of the freeway,the possibility of the disastrous of visibility such as the fog of the mission is increasing constantly,it is easy to cause the traffic accident,to the country and the people Caused incalculable losses.Therefore,the study of highway group fog monitoring and early warning is paid more and more attention by the state,so as to design a scientific,perfect and practical road group fog warning vehicle terminal system which can provide drivers with efficient and accurate real-time voice warning and effectively reduce fog and so on.Meteorological disasters caused by casualties and economic losses are the needs of social development.The rapid development of computer vision technology and risk assessment theory provides a powerful and convenient method for the study of warning and warning of group fog disaster.Therefore,the development of vehicle group fog early warning system based on video analysis technology has broad application prospects.In this paper,AM3517 development board is used as the hardware platform of vehicle intelligent terminal.Based on OpenCV computer vision library,combined with CCD sensor technology,embedded development technology and Android C ++ as development tools,this paper uses “qualitative judgment + quantitative discrimination” Koschmieder model and the Canny-Deriche filter,the Nagao filter and the region growing method are used to obtain the image's brightness distance curve and calculate the atmospheric extinction coefficient so as to obtain the weather visibility distance.Fisher discriminant analysis method and Logistic regression model gives the risk level and the probability of group fog,and finally real-time voice warning to remind the driver to reduce the speed to prevent traffic accidents.The main design work includes the functional structure design,the main interface design,the image acquisition based on Android Camera,the image preprocessing,the rapid area growth method,the corner detection and the meteorological visibility distance calculation for the mission fog early warning system software.The related software Interface and some of the key code.Mission fog early warning system testing and analysis results show that this article designed vehicle terminal performance testing and functional testing pass and achieve the desired goals,mission fog early warning system software can automatically detect the vehicle in front of the vehicle group fog state,and can Voice warning,with practicality.