License Plate State Recognition - Logo Matching and State Name String Classification - Issuing State Recognition Method

Sep 2, 2026

Vehicle license plates often include both alphanumeric characters and the issuing state or province. Robust state recognition adds valuable context for traffic management and improves segmentation and recognition of characters. Existing approaches relying only on state-name strings fail when the text is too small for low-definition cameras or is obscured by frames. This study introduces an issuing state recognition method combining adaptive unique logo matching with state-name string recognition, achieving higher accuracy across high- and low-definition video inputs.

LanguageEnglish

CategoryGeneral

FormatPDF

Pages8

Reading Time3 min

AccessFree to download

Abstract

Keywords

1 Introduction

Motivation and prior work

Proposed overview

2 (Method) State name string recognition

| Applied Mathematics & Information Sciences An International Journal | | | --- | --- | | [http://dx.doi.org/10.12785/amis/090239](http://dx.doi.org/10.12785/amis/090239)
License Plate State Recognition based on Logo Matching and State Name String Classiﬁcation
Zhongyan Liang1 ,∗, Sanyuan Zhang1 and Xiuzi Ye2
1 College of Computer Science and Technology, Zhejiang University, 310027 Hangzhou, P. R. China
2 College of Mathematics & Information Science, Wenzhou University, 325003, Wenzhou, P. R. China
Received: 10 Jun. 2014, Revised: 8 Sep. 2014, Accepted: 10 Sep. 2014
Published online: 1 Mar. 2015 | | | Abstract: In most countries, vehicle license plates contain both alphanumeric characters and the state/province of origin. State/province recognition of the license plate provides additional information to the trafﬁc management agency and guidance information, aiding in character segmentation and in recognition. Existing methods use the character string of the state name only. Unfortunately, in some countries, the state name is too small to be captured clearly by a low-deﬁnition video camera and may also be obstructed by the license plate frame. In this study, we analyze the characteristics of license plates and propose an issuing state recognition method based on adaptive unique logo matching (AULM) and state name string recognition (SNSR) . First, the AULM approach utilizes template matching to recognize the distinguishable logos. Second, if the logo is not located, the SNSR method is applied. The new method is compared with existing practices, and the experimental results show that the proposed method achieves higher accuracy and is more suitable when using high-or low-deﬁnition video images.
Keywords: state recognition, state name recognition, support vector machines, template matching | | | 1 Introduction
Recognition of the issuing state is the identiﬁcation of the attribute state using the information shown on the license plate [ 1] . Generally, the license plates contain cultural references or slogans associated with the issuing state in addition to the alphanumeric characters [2] . Usually, the state name string is shown along the top or bottom of the plate face. It may be difﬁcult to recognize the issuing state by the state name string alone, because of the size of the license plate frame, including advertisements inserted by the vehicle service center or the dealership [3], or the poor deﬁnition of the video cameras used to capture images. Moreover, the plate styles are not uniform. Most license plate recognition systems only report the license plate letters and numbers, without including the state information.
Researchers have studied a similar problem in some countries, i.e., attribution/address recognition. Because the format of license plate is different, the approaches to different countries are different. For example, in China, the ﬁrst Chinese character and English letter represent the province and city information, respectively. The license | plate recognition (LPR) system can report attributes by reading the main characters [4][5][6] . Similarly, in Germany, the 2 - 3 letter combination represents the region or city of registration [7] . For Japanese and Korean license plates, the address string can be used to recognize an attribute [ 1] .
Nevertheless, if the license plates are partially obstructed by its frame, the state name string will not be legible. Exactly extracting the name string from personalized license plates remains challenging (Fig. 1) .
The license plates used in this paper are captured mainly in the United States (US) . By analyzing the characteristics of license plates, some standard patterns, including the color schemes and logos, can be used to differentiate the license plates of different states or even different categories, such as a university, the army, or the navy [8] (Fig. 2) . We use an adaptive template-matching method to recognize the logos ﬁrst.

FAQs

Why is state/province recognition important for license plate processing?

State or province recognition provides additional information to traffic management agencies, supporting character segmentation and downstream recognition tasks.

What are the limitations of methods that use only the state name string?

The state name may be too small for low-definition cameras and can be obstructed by the license plate frame or poor video quality.

How does the proposed method recognize the issuing state when logos are present or missing?

It first applies adaptive unique logo matching to locate distinguishable logos; when the logo is not detected, it switches to state name string recognition using a support vector machine with PCA features.

Sep 2, 2026

Upload by Clementine