Research On All-weather Auxiliary Highway Traffic Technology And Design Of Support Platform

As the scale of China’s highways continues to expand,traffic accidents in this scenario are frequent growing.Thus,it is a significant step to vigorously develop smart highway for achieving safe driving,improving the efficiency of highway traffic and reducing the economic loss of highway.In the current boom of smart highway construction,quasi-all-weather passage is one of the key construction contents,and the integration of advanced technology to build a remote monitoring cloud platform is the main means.Remote monitoring of vehicles can provide early warning of abnormal vehicle conditions,promptly remind vehicle owners to maintain appropriate speed and focus on the road conditions ahead,so as to improve their traffic awareness and traffic capacity,reduce the incidence of traffic accidents.Ultimately,it plays a vital role in giving data support for future traffic disputes and facilitating the accurate division of responsibility.Therefore,this paper designs a highway all-weather passage support platform to remotely monitor the vehicles in transit on the highway.Meanwhile,in response to the most common problems on highways such as blurred video images in front of vehicles and difficulty in judging road conditions,which are mainly caused by rain and fog typical bad weather,this paper designs an image recovery algorithm based on weather classification and makes a distance judgment and speed limit value calculation for the recovered clear images.Firstly,an image restoration algorithm based on weather classification is designed to process the video image in front of the vehicle.Starting with weather classification algorithms,a weather classification algorithm based on convolutional neural networks is designed,which divides the weather into three types,including sunny,rainy,and foggy days.An improved dark channel fog removal algorithm based on non-mean filtering and a rain removal algorithm based on multi-channel convolutional neural network are designed.Through comparative verification,it is proved that the algorithm in this paper has good restoration effects for both rain and fog images.Then,a study of SC＿yolov5-based distance determination method is conducted.In comparison to the other four classical algorithms,the SC＿yolov5 algorithm is better suited for high precision and low latency usage situations thanks to the addition of the attention module.Through comparative experiments on target detection before and after restoration of rain and fog images,it can be concluded that image restoration can effectively improve the accuracy of target detection.By converting pixel coordinates into world coordinates,the real-time car distance is calculated,and a study of the experimental data reveals that the method’s average error is 6.3%,which can be used in practice.Then,the speed limits for rainy and foggy days are calculated.Driving too fast will raise the likelihood and seriousness of traffic accidents,according to research on the impact of vehicle speed on traffic safety.The visibility-based dark channel calculation method is presented for the calculation of the platform’s speed limit.Speed limit models are created for foggy and rainy days based on the parking sight distance model and road safety regulations,and a comparison table of advised speed limits for different weather situations is generated.Finally,the all-weather highway access support platform was designed and implemented.The overall architecture of the platform is designed,and the communication process between vehicle terminals and the platform is designed according to the actual monitoring requirements;then the data storage design is carried out according to the database design principles,and the security design is carried out from three aspects: user login,data storage and data transmission first;the detailed design and implementation of functional modules are carried out based on the requirement analysis;the image recovery algorithm,distance judgment method and speed limit calculation method are applied to the platform.The platform is used in conjunction with the distance computation method,speed limit calculation method,and picture recovery algorithm.The functional and performance tests of the platform were carried out,and the platform ran stably and can meet the design requirements.