Fake-licensed Car Recognition Based On License Plate Recognition And Points Marching Of Vehicle

With the vigorous development of the automotive industry, the number of cars isgrowing rapidly. The car is increasingly playing an important role in the transport andpeople’s daily life. The criminal activities which is involved in the car deck are alsoincreased.The cause of this situation includes legislation vacancy and managementshortcomings. Now, recognition in the domestic car deck rely on passive artificial way,which is not only time-consuming, but also a great waste of the police force, and can notobtain a better blow effect.The independent recognition system on the car deck is mainly reflected in:1.Vehicle identification and extraction. This part mainly studied on the backgrounddifference method to the vehicle identification and extraction. And the background modelupdating method is also studied in this part, which can lead a strong ability to adapt to anenvironment and time variation of the background update method, extract the targetrectangle of the target vehicle and short the time of the positioning plates.2.License plate location and character segmentation to identify. In this part, I mainlystudied in the the vehicle bounding rectangle image. According to the characteristics of theimage of the license plate, by the edge detection method to quickly locate the license plate,and in accordance with the license plate portion of the projection distributioncharacteristics of a candidate plate region a screening method; I also studied the methodfor correcting the tilt plate image, the structural features of the license plate character fastercharacter edge detection method. The study of license plate character recognition process isaccompany with the BP neural network.3.Vehicle matching feature points. In this part, I mainly analyzed the application ofvehicle identification based on the SIFT feature point matching. I also put a detailed studyof the establishment of the scale space of the image of the vehicle, the vehicle imagefeature vector generation and analysis of the matching algorithm, the faster choice of the BBF search strategy.