Research On Vehicle Object Contour Clearance And Velocity Measurement Based On Monocular Video

Intelligent transportation system is the mainstream of the future development, advanced traffic control and effective combination of computer science will be efficient serviced by human, which will greatly improves the processing efficiency of the regulatory and illegal accident. At present, the regulatory authorities is mainly by vehicle license plate which is recognized by the vehicle license plate automatic recognition system, as evidence of illegal to pursue. But for deliberately obscured license plate, license is not linked, licensed deck and other illegal acts, the existing license plate recognition system does not recognize to punishment by manual identification is time-consuming, which to solution to this problem has become a hot and difficult in intelligent transportation system.The content of this paper is the foundation work of distributed video vehicle unified spatiotemporal Association retrieval, provide important technical support for distributed video vehicle unified spatiotemporal Association retrieval. The main point of this paper are as follows:1 Video sample establishment and evaluate the quality of the video. For highway bayonet different perspectives, different road conditions, weather, and light constantly changing and other factors will affect the extraction of vehicle objects and further clear treatment. Therefore, establishment video sample based on different bayonet and evaluate, the video quality evaluation method based on moving target region ROI, corresponding video preprocessing will be performed by video quality evaluation result, so that the video quality evaluation effective associated video pretreatment.2 Extract the vehicle and clear it contour. Firstly, using the Gamma illumination compensation, Gaussian filtering, bilateral filtering, Laplacian sharpening, histogram equalization and other pretreatment methods, and then extract the target vehicle by mixed Gaussian background modeling method after video pretreated. In order to get a complete vehicle object by eliminate shadows with HSV-based shadow color space cancellation algorithm. Finally, by sequence inter-frame motion compensation method to repair contour and introduced the evaluation function of contour restoration, to control reasonable and effective of repair, ensure the contour of vehicle object clear and complete.3 speed of video based on lane. using standards information of highway lane to establish a speed lane area, calculate frames by of through vehicle in out of speed lane area and converse to time by video frame rate, Finally, calculate the vehicle speed.