License Plate Detection Using AdaBoost.pdf

License Plate Detection Using AdaBoost Louka Dlagnekov Department of Computer Science Engineering UC San Diego La Jolla, CA 92093-0114 Abstract License Plate Recognition (LPR) is a fairly well explored problem with many successful solutions. Though most of these solutions are reasonably fast, there can be increased benefits to making them even faster, such as multiple recog- nition stages in video frames in a real-time video stream for improved accuracy. The goal of this project is to evaluate how well object detection methods used in text extraction[1] and face detection [2] apply to the problem of LPR. A strong classifier is trained by the AdaBoost algorithm and is used to classify parts of an image within a search window as ei- ther license plate or non-license plate. 1. Introduction In any object recognition system, there are two major prob- lems that need to be solved – that of detecting the object in a scene and that of recognizing it. This project will mainly focus on the detection mechanism and rely on third-party OCR software for recognition. Most LPR systems employ detection methods such as corner template matching [3] and Hough transforms [4] [5] combined with various histogram- based methods. Viola and Jones have proposed very effi- cient object detection methods [2] that have been applied successfully to text extraction [1]. It is reasonable to pre- sume that these methods will also apply well to the problem of detecting license plates. This was, indeed, found to be the case. The methods used involve training a strong classifier using the AdaBoost algorithm. Over several rounds, Ad- aBoost selects the best performing weak classifier from a set of weak classifiers, each acting on a single feature, and, once trained, combines their respective votes in a weighted manner. This strong classifier is then applied to sub-regions of an image being scanned for likely license plate locations. An optimization introduced by Viola and Jones involves a cascade of str