Font Size: a A A

Video Detection Key Technology Research Of Traffic Flow Information In Urban Traffic Electrical Mechanical Systems

Posted on:2007-10-06Degree:MasterType:Thesis
Country:ChinaCandidate:F J HuFull Text:PDF
GTID:2132360185484651Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
Under the present traffic situation in Hang Zhou, this paper studies the problem of collecting traffic flow information using detecting and tracking of moving targets on the basis of video detection. The research work includes as follows: the research of using mixed Gauss distribution modeling background based on the gray, the comparability operator for gathering characteristics of moving targets and tracking multi-targets, and the information gathering problem of dynamic traffic flow using video technology.Firstly, this paper establishes an improved background absolute subtraction using Gaussian Mixture Models based on the luminance, by decoding and controlling the video based on two technologies of the Intel OpenCV graphics processing library and DirectShow streaming media processing, as well as reducing the noise of the image.Secondly, it deals with the binary image by mathematic morphologic method such as the corrosion and expansion of image after the image binarization to the results of motion detection with the Otsu method.Thirdly, it collects various parameters of moving targets, such as area, barycenter, the minimum external rectangle, and uses comparability operator for tracking multi-targets, which simplifies the calculation and can count the number and speed of the vehicles more precisely, solving...
Keywords/Search Tags:video detection, traffic flow, image processing, Gaussian mixture models, multi-object tracking
PDF Full Text Request
Related items