Vehicle License Plate Recognition Algorithm Based on the PowerPC Architecture
|School||Zhejiang University of Technology|
|Keywords||license plate recognition localization algorithm character segmentation character recognition powerpc|
With the continuous development of modern society, our intelligent traffic management system is also under rapid development. Smart license plate recognition technology as an important component of traffic management in the entire system architecture at an important position . License plate recognition system is a dedicated machine vision processing system , it is a car license plate on the image for a particular image, use of computer graphics and image processing and pattern recognition technology, to treat the collection vehicle detection image processing. Through digital image recognition algorithm running, the system can automatically and quickly and accurately identify the various characters on the license plate symbols, including numbers, letters and Chinese characters, and by the embedded device or a computer terminal, directly give the result of recognition. License plate processing system can make intelligent delivery vehicles into unmanned monitoring and management of real and intelligent level. Research on the license plate recognition system is of great significance in scientific research and engineering applications.This article describes a framework based on high-performance processing chip PowerPC core chip MPC5200 embedded license plate recognition system. System through the front end detection sensor to monitor vehicles entering the area to be monitored, and if a vehicle into the trigger image capture devices, capture the image after the video decoder chip through the video processing to the main processor port, and through license plate recognition image processing algorithm, according to license plate location algorithm, character segmentation, character recognition algorithms to capture images of license plate information is extracted, the result will be increased recognition to the terminal display and control of PC processing.Paper from the following aspects are discussed in detail:1. The system’s overall design: from the perspective of the overall design, designed the system function of each module, applications and design ideas, the overall grasp of the macro framework.2. License plate recognition algorithm: recognition algorithm consists of three main modules: license plate location algorithm, character segmentation, character recognition algorithm. This paper describes in detail the principle of the algorithm design ideas: using the SOBEL operator based on the license plate location algorithm; the connected components algorithm is applied to the character segmentation algorithm being; using BP neural network for character recognition algorithm.3. The system’s overall hardware design: the overall demand for the system to determine the overall design of hardware design, the schematic design and PCB board production.4. The system’s overall software design: increased use of the LINUX-based real-time operating system, license plate recognition algorithm will be packaged as a separate algorithm of nuclear units. Algorithm implementation including VC environment from the environment to the PowerPC transplantation and code optimization algorithms.