|
The license plate recognition system is one of the important research topic in computer vision and pattern recognition technology in the field of intelligent transportation applications. China has a growing number of cities have established a modern urban traffic control system, played a significant role to improve urban transportation to reduce traffic accidents. Development of license plate recognition system can greatly reduce human input, automatic monitoring of city vehicles. United States, Japan and other developed countries are relatively early design of intelligent traffic management system, and the application of sensors, cameras, sonar and other advanced technologies. Late start in China in this regard, intelligent traffic management, especially the intelligent analysis of the image is very urgent, therefore, this paper carried out on this aspect of the design is timely and of great significance. The financial support of the Natural Science Foundation of Anhui Provincial Education Department, with Tongling Blue Shield photoelectron Limited cooperation in the research and development of the system, and to participate in the research and development of on of Haibo Kang Security license plate recognition system. This paper from the overall system architecture, hardware platform, algorithm design, and software development in four areas of license plate recognition system design process. Why use TI's TMS320DM642 as processor system hardware, and describes the various modules in the TMS320DM642 processor hardware system, including image acquisition module, memory expansion module, Ethernet network interface; software design Research and Implementation of a license plate recognition algorithm, which is the focus of this work, including the license plate location, the correction of the angle of the image and character recognition; license plate positioning and image angle correction, a large number of experiments, and describes several The method of calculation of the angle, thus selected angle correction algorithm, when the characters in the binarized, a global threshold and a local threshold bersen algorithm combining binarization method suitable for application in license plate recognition. Program optimization section describes the development of fixed-point DSP in the CCS program should pay attention to the problem, and how to improve the efficiency of the implementation of the code; software development part of the software development environment CCS DSP software development processes, real-time low-level software DSP / BIOS use advantage of its drive model, based on RF5 reference framework program design. Hardware system video input and output digital signal acquisition program, the license plate recognition algorithm design can be achieved general recognition of license plate. Experiments and time limited, license plate recognition software, there are some inadequacies in strong sunlight or low light license plate low contrast, will be lost license plate. The hardware system using the DM642 processor, enabling the system to be applicable to a variety of highly complex image processing algorithms, and provides a high-performance platform for the development of the follow-up function, the system has a high practical value and broad application prospects .
|