| Mixed traffic flow is one of the main characteristics of urban traffic, Pedestrians and non-motor vehicles and motor vehicle constitute the main body of traffic. But researchers pay close attention to the study of vehicle as the key point of traffic monitoring and control, ignored pedestrians and non-motor vehicles. In this paper, we study the mixed traffic flow in real time detection and trajectory analysis based on video. The research objects are mixed traffic flow of vehicles and pedestrians in this paper, dive into the following aspects.(1) Summarize research status of the mixed traffic flow processing of home and international, analyzing defects of the processing methods for mixed traffic flow of real-time detection of tracking technology. On this basis, propose the mixed traffic flow real-time detection and tracking algorithm based on video processing, including:the target detection module, target tracking module and traffic parameter extraction module.(2) Put forward to apply a combination of methods of the gradient histogram (Histograms of Oriented Gradients, HOG) and support vector machine (Support Vector Machine, SVM) to complete object recognition and detection in complicated outdoor scene. Firstly, using the SVM classifier to achieve the goal of classification training for mixed traffic flow of vehicles and pedestrians, then using HOG feature extraction to achieve target detection.(3) Target tracking module uses the kalman filter, using the results of video images of the previous frame and current frame, estimate the location of the object in the next frame, so as to realize object tracking. Ensure the accurate estimation of target motion state.(4) On the basis of object detection and tracking, complete the trajectory analysis according to the trace of the moving target, obtain important traffic information such as flow rate, density and speed, etc, from a macro point accurately and effectively analyze motion behavior of target. |
PDF Full Text Request