Detection of vehicle pressure line based on machine learning

Intelligent transportation system (ITS) has been considered by scientists as the most effective way to solve the current urban transportation problems, and is also the mainstream direction of current and future transportation development. The premise of ITS is to obtain real-time traffic information, such as traffic flow. In this paper, we study the statistical method of traffic flow based on video detection technology in ITS. By analyzing the sample images, we obtain the basic characteristics of the video image of the moving vehicle, and select the image preprocessing method for this. At the same time, detect and extract the background of the moving vehicle video. According to the characteristics of the double yellow line area in the urban road image, the texture of the double yellow line area is increased by machine learning with the deformed Sobel operator, and the rough position of the double yellow line area in the image is segmented to obtain the accurate position of the double yellow line area. The change in the number of pixels in the virtual coil is counted to determine whether a vehicle is passing and whether the line is pressed, and the traffic flow and total traffic of each lane are counted respectively, and the vehicle type is determined by the length of time the vehicle passes through the virtual coil. It provides an effective solution to solve the current urban traffic problems, and can more accurately and effectively obtain real-time information on traffic roads.