Research On Vehicle License Plate Automatic Identification System

In the era of big data and smart city construction, vehicle license plate automatic recognition is the very important research topic and important application link in the field of intelligent transportation, widely used in the detection of illegal, vehicle automatic charge and other aspects. Identify the license plate number string from the actual scene, there is a lot of random interference, but also along with its own characteristics, so we require a series of processes to reduce these effects.In this paper, we first execute pretreatment for the vehicle background image obtained, and then exploit license plate location, license plate image segmentation, image character recognition process, to extract the license plate number string, and analyze the recognition performance. This paper also introduces and analyzes the other algorithms to support identification system.This article first arranges the composition and design of license plate recognition system, and describes the functions and workflow.License plate location is the whole key step in the recognition system, this paper first executes vehicle background image pretreatment, including gray, image enhancement and filtration, and then get the exact license plate image through edge detection.License plate image segmentation is first to combine the contrast stretching with the Bernsen algorithm to binarize the extracted license plate image, and then utilizing the improved Hough algorithm to correct the slant plate image, and finally adopting segmentation algorithm based on gray threshold to segment.Character recognition is the decisive part of the whole image recognition system, we use the improved BP neural network model as a recognition method to identify individual characters of image segmentation, to solve the slower convergence rate and local minima and other issues of the nonlinear identification process.Finally, the chapter analyzes and summarizes the recognition results.Finally, the paper summarizes the difficulties and need for improvement of the system, and subsequent studies are discussed.