The Application Of Lane Detection And Vehicle Detection And Tracking Technology In The Intelligent Traffic

The intelligent transportation system in the field of road traffic is one of the most important research areas today,whose purpose is to solve a series of traffic problems on daily lives.Based on the intelligent transportation system of Henan University,lane detection,vehicle detection and vehicle tracking are studied in this paper.The main works are as follows:1.The lane detection method based on Hough transform and K-means clustering is proposed.Firstly,the video image is resized,the region of interest is set,and the video image is denoised by Gaussian filter.Secondly,the edge is extracted by Sobel and Canny operators jointly.Then,the lines in image are detected by Hough transform.Finally,the lanes are detected out by using the K-means based road fitting method.2.A Gaussian mixture background model based and A Adaboost cascade classifier based vehicle detection method are realized,respectively.The Gaussian mixture background model based vehicle detection method firstly performs the foreground/background segmentation by Gaussian mixture model,and then collects the dynamic foreground contours after corrosion and expansion treatment.Finally,the real-time vehicle detection is finished after checking the contour area,height-width ratio,and duty ratio and so on.The Adaboost cascade classifier based vehicle detection method firstly collects the positive and negative samples offline.Then,the Adaboost cascade classifier is trained by using the Harr-Like feature of the vehicles.Finally,The vehicles in image are detected out by using the multi-scale detection method.3.A Kalman filtering based vehicle tracking method is realized.Firstly,the initial position of the vehicle target in the video image is given.Then,the measurement and predicted value of the Kalman filter are continuously updated.As a result,the vehicle can be tracked well.Moreover,the multi-vehicle selection tracking is also realized in this paper.Experimental results show that the proposed lane detection,vehicle detection,and vehicle tracking methods can meet the requirements of intelligent traffic monitoring system.