International Research Journal of Engineering and Technology (IRJET) Volume: 04 Issue: 05 | May -2017
www.irjet.net
e-ISSN: 2395 -0056 p-ISSN: 2395-0072
Survey on License Plate Localization using Genetic Algorithm and Temporal Redundancy Shreyas M1, Yashwan R1, S.Mahalakshmi2 1Student,
Department of ISE, BMSIT&M Dept of ISE, BMSIT&M Bangalore, Karnataka, India
2Asst.Prof,
---------------------------------------------------------------------------***-----------------------------------------------------------------------------Abstract: License Plate Recognition plays an important 2. PROPOSED SYSTEM role in real world applications such as automatic toll collection, access control in private parking lots, stolen 2.1 Temporal Redundancy Approach vehicles identification and traffic surveillance. Recognition of an on-road vehicle using its license plate is an ALPR based on temporal redundancy involves: important task performed by several intelligent transportation systems around the world. This task is 2.1.1 ALPR Pipeline known as Automatic License Plate Recognition. This paper 2.1.2 Temporal Redundancy Aggregation is a survey of license plate Recognition based on temporal 2.1.3 Vehicle Appearance Classification redundancy and genetic algorithm integrated with image 2.1.4 Tree Based Query thresholding. ALPR using temporal redundancy discusses mainly about temporal redundancy approach along with 2.1.1 ALPR Pipeline two post processing techniques. Genetic algorithm on the other hand, executed the license plate region detection of We first detect the vehicle and then its license plate, the digital image which depends on the set-level of the located inside the vehicle patch. To solve both tasks, we image threshold values obtained. employ a sliding window approach composed of a classifier based on Support Vector Machines (SVM) and Key Words: Automatic license plate recognition, vehicle Histograms of Oriented Gradient (HOG) as feature classification, Genetic Algorithm, GRM, CCAT, and descriptors. Once the license plate has been located, we Image Thresholding need to segment the image into multiple patches containing license plate characters (LPCS). . In this approach, instead of using a single threshold to perform 1. INTRODUCTION license plate binarization using the Otsu method, we consider a set of different values. It is believed that there are currently more than half a billion cars on the roads worldwide. All those vehicles have their vehicle identification number as their primary identifier which is mounted onto its body and no vehicle without properly mounted, well visible and well readable license plate should run on the roads. With increasing number of vehicles on roads, it has become difficult to manage traffic rules for smooth traffic flow. Vehicles in each country have unique license number which can be seen on the license plate. This number helps to distinguish one vehicle from other. An automated system can be implemented to identify the Fig. 1: Samples of the license plate considering different thresholds, 5 vehicle number plate and extract the character from the and 10 on the top images and 20 and 30 on the bottom images. region containing the license plate. The number recognised from the license plate can be used to retrieve 2.1.2 Temporal Redundancy Aggregation information about the vehicle and its owner which helps in further processing. We combine the individual recognition results using two main approaches: (i) majority voting and (ii) average of the classifier confidence. While the former takes all predictions for each frame and assumes that the most predicted character for every license plate position is the correct, the latter averages the classes by minimizing the intra-class variance. Š 2017, IRJET
|
Impact Factor value: 5.181
|
ISO 9001:2008 Certified Journal
| Page 1501