Mathematical Morphology Character Segmentation Algorithm Apply And Research In Vehicle License Plate Recognition System

Vehicle license plate recognition system is an important subject in the research field of intelligent traffic system. License plate recognition through to capture the image of the passing vehicles,then the license plate part from a vehicle image will be located and analyzed the information to identify the license plate.Comparing the information with license plate data in the database can be a good vehicle monitoring and management. The license plate recognition system is widely used in highway toll management system,parking fee management system,residential property management,road management system,etc.It provide a strong technical support for the implementation of the vehicle odd-even traffic control measures to help relieve the pressure on urban traffic. Therefore, the research of license plate recognition technology not only has important practical significance, but also has a vast market and great prospects for commercial applications.The system theoretical research of vehicle license plate recognition include license plate positioning, character segmentation, character recognition, and the character segmentation is the focus of this thesis. Digital image processing technology is the main theoretical basis for the research license plate recognition system, therefore,the paper summarizes the basic knowledge of digital image processing technology and analyzes the license plate image preprocessing, the license plate positioning, character segmentation, character recognition technology simply.This paper focuses on the research of the license plate character segmentation algorithm based on mathematical morphology,which is the emphasis of the thesis study. Introducing the theory of mathematical morphology,which is an image processing and analysis of emerging disciplines and including dilation,erosion,opening operation and closing operation technical parts. Mathematical morphology processes image efficiently and programms simply, in view of this, the paper presents a license plate character segmentation algorithm based on the horizontal and vertical gray opening operation.Usually,after the license plate positioning,it often has a little angle of inclination and include some frame and rivets,which will cause serious interference with character segmentation. Firstly, using the Hough transform to detect license plates tilt angle and correcting plates, then removing the license plate frame and rivets by the horizontal and vertical grayscale opening operation, finally, through the traditional projection method and prior knowledge of the license plate segmentation between characters. By comparing with the license plate character segmentation which the frame has not been removed completely, proved by mathematical morphology algorithm of this paper removing license plate frame effectively,the effect of the segmentation of characters is improved.Finally,the paper introduces segmentation algorithm of license plate character based on the the DM642DSP development platform.It introduces the TI DSP chip and technology simply and analyzes two hardware modules of the image acquisition and image display in the DSP development board.