Vehicle Detection And License Plate Recognition Based On Video Image

License plate recognition system is an important part of Intelligent Transportation System(ITS).License plate recognition has been widely used in urban traffic management,parking lot management,vehicle access,highway toll collection and so on.For the urban road traffic bayonet system,the efficient detection of the vehicle,the identification of the license plate is to help crack the escape,vehicle violations and vehicle loss and other effective means of the incident.In the license plate recognition system,the accurate detection of the running vehicle and the license plate location has an important effect on the recognition of the license plate character.If the license plate can not be accurately positioned,character recognition will not be carried out.Therefore,this paper focuses on the further research of vehicle detection and license plate location,then put forward a new vehicle detection and license plate location license plate location method,and the method was realized by programming in the Matlab.The experimental results show the effectiveness of the proposed method.The main contents of this paper include the following aspects:Detection of moving vehicles.The advantages and disadvantages of some frequently used foreground extraction methods were analyzed.Three different methods of vehicle extraction were programmed on the Matlab.Based on the analysis results,the background model of the background difference method was selected to detect the vehicle.According to the actual needs,the vehicle is detected by the average background model of the background subtraction method and the gradient image.Finally,the target vehicle is extracted by morphological method and binarization method.License plate location.Some methods commonly used in license plate location were briefly introduced,and analyzed their advantages and disadvantages.Finally,the shortcomings of some license plate location methods were verified by experiments,for example,the edge detection method has a big error for the vehicle license plate location with poor image quality.According to these shortcomings,a license plate location method combining edge detection and color feature was proposed,and the reliability of the method was verified by experiments.License plate tilt correction and character segmentation.After successfully locating the license plate,it is necessary to prepare for the license plate recognition.The positioned license plates often exist problems with tilt and positioning insufficient accuracy.According to these problems,some algorithms were used to correct the tilt plate,and the corrected license plate was used to locate the license plate and segment the license plate character according to the vertical projection method.Characters recognition.The existing license plate characters recognition algorithm were analyzed,and then the appropriate algorithm was used to identify the license plate characters.