| In recent years,China chemical industry has maintained a great development trend with the continuous development of the national economy,which has also increased the demand for the transportation of hazardous chemicals.Because of the long driving distance of hazardous chemical transport vehicles,the main factors for accidents are the driver’s fatigue operation,non-compliance with traffic rules,loss of attention etc.If the influence of human factors is minimized as much as possible,unexpected traffic accidents can be avoided.The vision-based front vehicle collision avoidance warning technology is an important research direction in the field of active safety of hazardous chemical transport vehicles.Especially,fast and accurate detection of the front vehicle and establishment of a stable and reliable safety distance model are two difficulties that this technology needs to solve urgently.For this reason,this paper proposes an early warning algorithm for highway anti-collision of hazardous chemical transport vehicles based on vehicle-road vision collaboration.Firstly,a lane line detection algorithm with multi-feature fusion proposed in this paper is used to detect lane lines in day and night environments.Then,the TF-YOLO(Tiny Fast Yolo Look Only Once)algorithm is proposed to detect the vehicle in front in real time and obtain the position information of the vehicle.Finally,the RCSZ(Real-time Calculation of Safety Zone)algorithm is proposed to construct a safe distance model to form an early warning safety zone in front of the current vehicle.According to the position of the preceding vehicle in the image and the constructed safe distance model,it is possible to deal with rear-end collisions.Experiments show that the lane line detection algorithm based on MFF(MultiFeature Fusion)proposed in this paper has good robustness,and is less affected by noise,the interference of yellow lane lines,shadows and strong lights.The proposed TF-YOLO algorithm can meet the detection of the preceding vehicle in a high-speed driving environment,with an accuracy rate of 98.04%.The early warning algorithm can provide lateral and forward early warning.Compared with the traditional anticollision warning methods of vehicle ultrasonic,radar or laser ranging,it has stronger applicability and stability,high warning accuracy,and can reduce the incidence of rear-end collisions during transportation. |
