Design And Implementation Of System For Plate License Detection And Recognition Based On DM8127

The license plate recognition system is the core part of intelligent transportation,and it is widely used in the fields of highway illegal vehicle monitoring,highway vehicle flow control,highway license plate recognition fees,and parking lot vehicle management.After years of research,this technology has made great progress.However,due to the complex factors in the real environment,the detection and recognition of license plates still has important research value,and with the advent of the era of big data,the data storage has grown in a geometric progression,which is also a great challenge for the use of information in intelligent transportation.This thesis mainly is studied a smart front-end license plate detection and recognition system based on DM8127.Combined with the smart front-end development environment based on DM8127,firstly,it uses OpenCV on PC to realize a license plate detection algorithm and license plate character recognition algorithm in a complex natural environment,and then realizes the license plate recognition system in the smart front end.The main work of this thesis includes the following aspects:(1)A location algorithm based on SVM which takes the color features and edge features of the license plate is designed.The license plate location method based on color features can be used to extract the license plate in a complex background.Firstly,the color information of the license plate is converted to the HSV color space,and then the license plate image can be quickly located from the fixed background color of the license plate;Based on the edge-based license plate location method,a secondary sobel edge detection method was designed to detect the license plate in complex environments such as scale changes and illumination changes.(2)A segmentation algorithm based on contour method and license plate's prior knowledge is designed.The contour method is used by this algorithm to obtain numbers and letters and to solve the problem of disconnection of Chinese characters combined with the prior knowledge of license plates.A BP neural network character recognition algorithm which takes the vector of the histogram and low-rate graph of characters as the feature is designed.(3)An intelligent front-end license plate detection and identification system based on DM8127 is built.The video acquisition module,front-end analysis module,and interface display module of the smart front-end system based on DM8127 are analyzed firstly.Then the license plate detection and recognition algorithm library which is based on the DM8127 intelligent front-end platform is designed.The algorithm library includes machine vision library layer,algorithm dependent library layer,and algorithm application layer.According to the actual platform conditions of the intelligent front-end monitoring system,the parameters of the algorithm are adjusted.The experimental results show that license plate can be detected by the intelligent front-end license plate detection and recognition system constructed in this thesis quickly and effectively,and the recognition accuracy is above 90%.