Font Size: a A A

Passenger Count Method Based On Detection And Tracking In Intelligent Video Monitoring System

Posted on:2015-12-22Degree:MasterType:Thesis
Country:ChinaCandidate:L XieFull Text:PDF
GTID:2322330485493697Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
Passenger count is the basis for the operation of intelligent video monitoring system. It can get the number of passengers get on and off and maintain the order and security of buses. With the rapid developing of image processing and pedestrian detection and tracking technology, passenger count method based on bus video monitoring system has become a hot topic. However, due to the complex environment of buses, vibration of video, trapezoidal distortion of images and the barrier among passengers, a real-time and accurate passenger count program do not exist.To solve these problems, the following technologies are proposed. Firstly, a video stabilization technology was proposed to solve the problem of vibration. Secondly, a trapezoidal correction algorithm was implemented. Thirdly, a passenger detection and tracking algorithm was presented to judge the direction of movement and calculate the passenger count.Firstly, this algorithm selected two sub-blocks in the upper corners to obtain the local motion estimation, and took the average value of two local motion vectors as the motion compensation vector, reducing the video vibration. Secondly, it achieved trapezoidal correction with projective transformation matrix and bilinear interpolation algorithm. Thirdly, it used adaptive threshold background subtraction method to achieve passenger segmentation, which can eliminate the interference in the background. Then it used Histogram of Oriented Gradient(HOG) feature of head-shoulder to train Support Vector Machine(SVM) classifiers. It extracted Speeded-Up Robust Feature(SURF) of target region to achieve passenger tracking, which can apply to the case of unstable illumination. Finally, this method used counting line to judge whether there were passengers getting on or off. It judged the direction of movement by detecting and tracking passengers, and then calculated the passenger flow at each site and counted up the total number of passengers.
Keywords/Search Tags:passenger count, video stabilization, trapezoidal correction, target detection, target tracking, HOG, SURF
PDF Full Text Request
Related items