Some Technical Research And Implementation For Vehicle License Plate Recognition

Vehicle license plate recognition is the use of computer vision, image processing and pattern recognition and statistics and a variety of ways and means to make the plate pretreatment, positioning,segmentation and recognition of vehicle license plate recognition technology; basically include image acquisition, image preprocessing, license plate location, license plate character segmentation and recognition of several important processes.Intelligent traffic system as an important technology in vehicle license plate recognition technology, in road monitoring, intelligent car park management and toll collection system and so on have important application; a large number of domestic and foreign researchers have been working on the license plate recognition technology and system research, and has achieved a number of good results; overall, vehicle license plate recognition technology matures; however this technology in license plate pretreatment, positioning and segmentation technology still has a lot of improving space, intelligent transportation system is gradually built and improved, the overall recognition rate remains to be further improved. Especially in multiple lanes, more complex environment, how to improve and overcome these negative factors to the recognition rate brought negative effect,will become a current license plate recognition in the field of research in a very important topic.Article on process pretreatment technology of vehicle license plate, license plate recognition and vehicle license plate character segmentation of license plate location technology for the study, improvement and implementation. Study on pretreatment technology in vehicle license plate, presented a modified license plate image binarization method:this method not only license plate image background and goals are very obvious, image reduces the need to handle a large amount of land in, at the same time be able to effectively get rid of the noise interference, for creating good conditions for increase of recognition accuracy; Vehicle license plate locating links, combined with the previous multiple positioning method based on gray level is made on the basis of this figure and aspect ratio are two important factors of vehicle license plate location, And is presented after the binarization using colored dots on the search to further improve the accuracy of vehicle license plate location; Segmentation of license plate, this article summarizes several typical split of ideas and technologies, as well as an improved method of vertical projection. In the hardware support system, this paper introduces the DM642as the core processor. The TMS320DM642chip is based on the TI company’s second generation high performance, advanced VLIW (very long instruction word) structure, with strong processing ability, high flexibility and programmability. At the same time, the peripheral complete with integrated audio, video and network communication equipment and interface, especially for digital video image processing applications. This paper also study and do some algorithm transplantation, transplantation is the so-called algorithm according to program operating platform, the program corresponding to different parts to modify, adapt to run platform. TI launched the DSP advanced language (C, C++compiler), the DSP chip software can be directly written in a high-level language, which makes the program modification and transplantation becomes simple; the related procedure is from Visual C++platform transplantation into the CCS platform, this paper extraction of VC++applications and algorithms related to the code, place it in the CCS, and then combined with the embedded DM642software platform structure, for its consistent with CCS operating platform of the amendment, thereby establishing the license plate recognition system code. In short, this article in the license plate pre-processing, vehicle license plate location and segmentation technology was improved, and the experimental results, and achieved satisfactory results; at the same time a partial correlation algorithm transplantation.