Extraction Of Traffic Parameters Based On Detecting Traffic Waves

With the rapid increase of number of motor vehicles, the traffic congestion on urban road has become increasingly severe. The improved traffic control strategies are needed urgently, and the accurately acquisition of traffic parameters is the premise of optimizing traffic signal control strategies. Therefore, the improvement of traffic parameters acquisition based on advanced information technology is of great significance to improve the city’s transportation situation.In recent years, based on machine vision, many scholars have done research on traffic parameters detection. The use of computer and image processing technology to get the traffic parameters has become a hot spot. It is a challenge to get high precision parameters by video processing technologies during peak hours, for the serious vehicle queue. We made researches on the method of acquiring traffic parameters based on detecting traffic waves at intersections. This paper in-depth study in the view of the above problems, and the main work and contributions include the following aspects.1. First, we introduce the method of manually recorded traffic wave value based on artificial calibration. Using the intelligent software VIPER to obtain the pixel coordinates of vehicles in video images. Adopting the pinhole camera model to convert the pixel coordinates into world coordinates of road plane. According to the vehicle in time position to determine the queue length and stop delay, and etc. Based on the break points in the vehicle motion curve, the traffic wave cure is fitted. The manually recorded values are used as real values2. Introducing three traffic wave position detection methods, including flexible window to detect queue length via a single camera, tracking results of two cameras at decision-level and pixel-level. We use experiments to take the manually recorded data comparing with results detected by those three traffic wave detection algorithms, and establish evaluation index to analyse each performance.3. Carrying out a research on obtaining traffic parameters based on the traffic wave detection. The parameters include queue length, stop delay, velocity and etc. there are some mathematical relationship between parameters and traffic wa ve position, and based on the relationship to deduce the calculating method of corresponding parameters. The calculation of average stop delay is determined by the queues number in a certain cycle.4. Development and improvement of manual calibration data export and parameters extraction system software. Based on the experiment, acquiring multi segments of traffic video under morning and evening peak hours respectively, and adopting related algorithm to obtain the queue length and stop delay. We compare the calculated values with manually recorded value to evaluate the performance of the method, and then make conclusions.