A system for licence plate recognition using a hierarchically combined classifier. (14th March 2011)

• Record Type: Journal Article
• Authors: Zheng, Lihong; He, Xiangjian; Wu, Qiang; Samali, Bijan
• Is Part Of: International journal of intelligent systems technologies and applications. Volume 10:Number 2(2011)
• Page Start: 189
• Page End: 202
• Publication Date: 2011-03-14
• Subjects: licence plate recognition -- classification tree -- hierarchically combined classifier -- licence detection -- character segmentation -- character recognition -- car number plates -- licence plates -- inductive learning -- support vector machines -- SVM; Artificial intelligence -- Periodicals; Intelligent control systems -- Periodicals; 006.3
• Languages: English
• ISSNs: 1740-8865

Abstract

In a real time, automatic licence plate recognition system, licence detection, character segmentation and character recognition are three important components. All these three components generally require high accuracy and fast recognition speed to process. In this paper, general processing steps for license plate recognition (LPR) are addressed. After three types of combined classifiers are introduced and compared, a hierarchically combined classifier is designed based on an inductive learning-based method and an support vector machine (SVM)-based classification. This approach employs the inductive learning-based method to roughly divide all classes into smaller groups. Then, the SVM approach is used for character classification in individual groups. Having obtained a collection of samples of characters in advance from licence plates after licence detection and character segmentation steps, some known samples are available for training. After the training process, the inductive learning rules are extracted for rough classification and the parameters used for SVM-based classification are obtained. Then, a classification tree is constructed for next fast training and testing processes based on SVMs. The experimental results show that the hierarchically combined classifier is better than either the inductive learning-based classification or the SVM-based classification with a lower error rate and a faster processing speed.