Design And Implementation Of Embedded Vehicle License Plate Recognition System

With the high-speed development of social economy and the development of road traffic in our country, intelligent traffic has become a focus in the study of science and technology. As the core component of intelligent transportation, license plate recognition plays an important role in the intelligent traffic management. Aim at large volume, high cost, poor flexibility, convenient installation of the traditional license plate recognition, this paper studies the system based on the embedded ARM, through the use of real-time operating system Linux to realize video collection, image processing, document storage, and use the embedded Qt to realize the friendly human-machine interface.At first, this paper introduces the relevant background and research status of the intelligent transportation and license plate recognition, then introduces the overall design scheme and identification process, including license plate localization, character segment and character recognition. In license plate recognition algorithm research of this paper, according to the geometrical and gray level characteristics of the license plate, we adopt a hierarchical approach to localization the plate; and on the basis of image tilt correction, an improved projection method is used to split character; Finally we use an improved neural network to identify the characters. The results show that the algorithm can meet the requirements of speed and accuracy. In the aspect of system design include the establishment of the Embedded Linux system, the cutting and transplant of kernel, as well as set up the cross compiling environment, using Qt/Embedded and OpenCV image processing function library to write the software and successfully completed the recognition algorithm.Embedded license plate recognition system is convenient, flexible, adaptable and able to suffice for many practical applications. After experimental test, the system can complete the image acquisition and the realization of the algorithm and has high recognition speed, high stability, low consumption and a good human-computer interaction function.