Research On License Plate Recognition Method For Traffic Surveillance System

License plate recognition (LPR) is the crucial component of the Intelligent Transport System (ITS). LPR plays important role in transport application such as travel time management, parking lot traffic management, speed limit enforcement, red light violation, identification of the stolen car etc. Lots of work has been done but there is still room for the improvement such as the license plate positioning ratio, character recognition ratio and so forth. As a result, the research intends to focus on this area in anticipation that the recognition accuracy of the license plate recognition system can be further enhanced.In this thesis, innovative methods are proposed for LPR that is targeted to solve the inherited issues. Modified and effective method of license plate area location is presented which is based on the morphological and image projection technology. At first Robert edge is used to detect the vertical edge contained in the license plate. Then apply morphological operations to find the exact candidate of license plate and remove extra region which doesn't belongs to plate. At last extract the plate region using horizontal and vertical image projection. Character segmentation is done using the prior knowledge of license plate and region growing algorithm. In proposed methodology at first, contrast stretching is performed to enhance the character region. Then to remove plate boundary, frame and rivet, row and column scan statistics algorithm has been applied. At last the candidate character is segmented by using region growing algorithm. This study focus on backpropagation neural networks learning paradigms and network framework combining normalization, parallel dispersed and reverse logical thinking concept to propose reverse logical artificial neural network (RLANN) for the character recognition. The proposed algorithm was implemented on MATLAB 9.1. Many samples are taken in different environment. The experiments indicate that it is feasible to adopt this algorithm in LPR to achieve better accuracy and adaptability.