Research Of The Technology Of License Plate Recognition

With the economic and social development, the number of cars in China isincreasing, we have paid more attention on traffic intelligent management, and how toimprove the traffic intelligent management will be a serious challenge. As thepowerful weapon in realizing intelligent management, the license plate recognitionsystem is the indispensable key part, it has become a study hotspot gradually in recentyears. The license plate recognition management system is an intelligent trafficmanagement which is mainly based on the combination of many techniques such asdigital image processing, pattern recognition and artificial intelligence technology. Itsmain process can be divided into three steps: license plate localization, license platecharacter segment and license plate recognition. Based on the analysis of the existingresearch on the theoretical basis, we should finish the following works which aimed atthe three key technologies in the license plate recognition system.1. Analysis the common use in the license plate localization algorithm today afterpreprocess the image of license plate localization, using the Canny operator to get theclearer result of license edge detection, and finally, use the characters which based onthe shape and the gray-scales to locate the license plate.2. In the process of segmentation algorithm of character, we correct the image ofthe license with linear transform algorithm, and then we use the gray-level projectionmethod to segment the characters of the license plate.3. In the process of recognizing the license plate, we use the method which basedon Hausdorff distance: first of all, improve the Hausdorff distance method, thennormalizing and refining the identification characters. After that, calculate thematching degree between the unknown and the standard modals. At last, deal with therecognition of license plate operation. According to verification test, such kind ofalgorithm can improve recognition rate in a certain extent.