Design Of Embedded License Plate Recognition Based On ARM

License plate recognition system is mainly composed of following modules: vehicle imageacquisition, image pre-processing, license plate location, character segmentation and characterrecognition.It’s not only the key component of modern intelligent transportation systems,but also ahot point of research in image processing,pattern recognition and artificial intelligence，So its relatedtechnologies are attended by more and more scholars and research institutions. And with the widelyapplication of the embedded systems in various industries,research and development in the field ofcomputers has become a hot spot.Based on this, the paper makes a deep study on the license platerecognition technologies, and the embedded license plate recognition system based on ARM isdesigned.The paper first of all introduces the background, significance and development trends of licenseplate recognition technologies at home and abroad.And then, a total solution of embedded licenseplate recognition system based on ARM is stated,which includes the design of hardware platformswith S3C2440at its core,the construction of embedded Linux platform,the transplantation of USBcamera driver, the programming of image acquisition based on V4L2and transmission and thedesign of license plate recognition software.At last,the simulation experiment of the algorithms iscarried out on the PC through Visual Studio.The algorithms of all steps for implementing a complete software include image preprocessing,license plate location，character segmentation and character recognition，and these are designed inthis thesis.In order to improve the image quality, pretreatment was performed before the license platecharacter location.In the license plate locating part, the license plate area is extracted based on thepriori knowledge like license plate size and the differences of texture features between the licenseplate area and the background area,and then the text area of the plate is extracted.After eliminatingthe background noise and the influence caused by frame of license plate and its rivets. In thecharacters segmentation system,based on the structure and the prior knowledge of the licenseplate,character segmentation algorithm which based on projection is used and the result proves thatthe method is effective, it establishes the foundation for character recognition.At last, the method offeature point matching is used to identify the character.In the last part of the thesis, the main content of the design is given,it also points out the defectsof the system and gives the directions for future work.