Research And Implementation Of License Plate Image Processing In Complex Environment

As a unique information of a car,vehicle license plate is the most effective way to identify a car,and then to make a unified intelligent management of all kinds of vehicles.As one of the core contents,license plate recognition technology has also been widely attented.At present,the technology has been widely used in high-speed intersection,self parking toll gate,the loss of vehicle tracking.But the existing license plate recognition system is suitable for the single situation,the recognition rate is not high in complex scene generally.Therefore,in this paper,we take the complex environment as the main research background,the complex environment and complex light of the vehicle license plate recognition technology are in-depth studied.In this paper,the method of license plate location based on region enlargement is used to adapt to the complex environment.First of all,combined the traditional edge detection methods with the morphological processing,the processing results of the noise is small and the marginal information is clear and completed;and then;as a result of two times of rough localization based on the iterative fusion of the connected domain and the multi-factor-decision,determine the most possible license plate area and remove the pseudo license;finally,aiming at the phenomenon that the angle of inclination of the license plate is large and the lost of the useful information of the license plate,this paper puts forward the method of area enlargement for license plate location.After the completion of the license plate image location,the license plate image is further processing for complex light.In order to solve the problem of uneven illumination images often appear in: the boundary of the character and background is easily confused,the lost of detail information,the noise is large,the light changes suddenly and scheduling problem.This paper proposed an image enhancement algorithm under complex illumination based on improved Retinex theory,the experimental results show the feasibility of the method from two aspects: subjective visual effect and objective indicators,and fully verify the feasibility of the method,and then achieved final results treated by various methods and processing by the self-adaption binaryzation,fully embodies this method has the advantages of high clarity,clear details and low noise.Finally,practiced and verified by experiments,through the analysis of experimental results,this paper proved that the method of license plate image processing technology under complicated environment is feasible,and compared the processing results with other methods,protruding the superiority of the method in this paper,finally pointed out the problems in the experiment and put forward the improvement measures.