Detection and Recognition of Multiple License Plate From Still Images

Detection and Recognition of Multiple License Plate From Still Images

Table of Contents


License Plate Recognition is the most efficient and cost-effective technique used for vehicle identification purposes. Automatic license plate recognition (ALPR) is used for finding the location of the number plate. These approaches and techniques vary based on conditions like image quality, the car at fixed positions, conditions of lights, single image, etc. It should also be able to cope with the variations in license plates from different nations and states. The approach should also be able to work seamlessly with the number of characters varying in plates or sizes of the plates in the captured images. We mainly focus on the detection and recognition of multiple car license plates from a single frame. The proposed system consists of two steps: plate number detection and recognition. In the plate detection part, we apply for both Spanish and Indian license plates. In our test case, we will be working with number plates from Spain. Three different license plates which differ from one another in their size and shape. In plate number detection phase the license plate is detected from the captured image and then in the second phase, the segmented plate is passed to plate recognition that makes to determine the characters and numbers.


Automatic number plate recognition (ANPR) can be used to store and process license plate images captured by cameras with a high rate of accuracy and efficiency. In ALPR we can enact different techniques based on varying conditions like image quality, fixed car positions, and multiple plates extraction. The ever increasing vehicle count on our roads has hindered the smooth flow of traffic. It finds great use in managing real-life applications such as border control, parking, motorway road tolling, journey time measurement, access control, road traffic monitoring, etc.

A vehicle in a country is distinguishably identified by a unique alphanumeric number, which will be depicted on its license plate. Systems commonly use IR cameras to take the picture. Due to a change or a distinct form in color, texture size, shape, and position of plate regions in such images the localization of plate regions is a challenging task. ALPR system completes the entire process passing through different stages. The stage is based on some features of the captured image to extract the license plate from the image. In the next stage by projecting their color information we can segment the license plate. After that characters are extracted. Finally, recognition using template matching is done.

Author Keywords

  • Number plate identification,
  • license plate recognition (LPR),
  • license plate segmentation,
  • edge detection.

IEEE Keywords

  • Licenses,
  • Image segmentation,
  • Image edge detection,
  • Automobiles,
  • Cameras,
  • Support vector machines,
  • Classification algorithms


cameras, specified car positions, limited vehicle speed, and light conditions designated routes, etc. We are mainly focusing on the detection and recognition of multiple car license plates from a single frame. Lp detection based upon combinations of mathematical morphology features and edge statistics produce high standard results [1] [2] [3]. Edge-based methods are not much used for complex images, for the reason that it is too sensitive to the unwanted edges [2]. Mathematical morphology [3] [4] that is based on nonlinear neighborhood operations are used.


Following are the contributions of this study,

  • Efficient car license plate detection and recognition system.
  • Detect and recognize multiple car number plates in a single camera frame.
  • Licenses with complex backgrounds are tracked correctly.

In [5], Panahi, Rahim, and Iman proposed an accurate online system for ANPR. Three main parts for this process: plate detection followed by character segmentation finally character recognition. They used a revised version of algorithm RANSAC for the license plate localization process. Different data sets were created Crossroad dataset and Highway dataset and were used for training and testing. The crossroad data set were collected from different crossroads. We obtained highway data set from highways and streets. Plate segmentation was done, input a grayscale image, applied a global/adaptive threshold then divided M*N blocks. After locating the plate region the characters are processed then, the output line is given to the character recognition part of the RANSAC. After that to predict whether a component is a character or not they use a two-class SVM. The framework accomplishes 98.7%, 99.2%, and 97.6% for three stages. The system is not language-dependent, achieved 97% overall accuracy.

Related Techniques

In [6], the image processing technique established on the number plate recognition (NPR) system, to identify vehicles by applying neural networks and image co-relation. Extracting license plate from lacking brightness and less brightness image obtained recognition rate of 96.64%. In [7], detection is used to identify vehicles, using image co-relation and neural networks. Different novel approaches have been presented to improve the results. Pattern recognition is divided into two broad categories: recognizing abstract items and recognizing concrete items. They were used a multi- thresholding and neural pattern recognition (NPR) techniques combined with artificial neural networks. They obtained a recognition rate of 98.40%.

In [8], an idea based on SVM recognition. Support vector machines (SVMs) consist of a set of related supervised learning systems. Two stages have been used for SVMs multiclass classification “one against all” and “one against one”. First method “one against all” (OVA) [9][10], binary classifiers sets were used. One against one ‘is applied to each pair of classes. They input 315-dimensional feature vectors into SVMs for training. They performed procedures on 180 images of license plates under different conditions. In [11], for plate detection, proposed several algorithms. Proposed canny edge-detection used based on optimization. The algorithm is based on clear and simple steps, six to be precise. Firstly read and resize the image, convert a grayscale image followed by a complement of the image was received and the edges found. Using filters image converted to small objects after that separate the image into objects. Finally recognition of the plate.


This paper presents a successful and quick process for detecting multiple license plates both Indian and Spain. Automatic license plate recognition (ALPR) is used for location detection of the number plate. The advantage of our proposed method on multiple plates is its high accuracy in the plate detection part. The proposed method detects multiple car number plates in a single-camera frame performed correctly. A license with a complex background is tracked correctly and obtained a good result.


We would like to give the first honors to God who gave the wisdom and knowledge to complete the work. We thank the chairman and the management of SSET for providing the infrastructure without which it would not have been possible to complete this paperwork and also we thanking our coordinator who gave valuable support to us.

About KSRA

The Kavian Scientific Research Association (KSRA) is a non-profit research organization to provide research / educational services in December 2013. The members of the community had formed a virtual group on the Viber social network. The core of the Kavian Scientific Association was formed with these members as founders. These individuals, led by Professor Siavosh Kaviani, decided to launch a scientific / research association with an emphasis on education.

KSRA research association, as a non-profit research firm, is committed to providing research services in the field of knowledge. The main beneficiaries of this association are public or private knowledge-based companies, students, researchers, researchers, professors, universities, and industrial and semi-industrial centers around the world.

Our main services Based on Education for all Spectrum people in the world. We want to make an integration between researches and educations. We believe education is the main right of Human beings. So our services should be concentrated on inclusive education.

The KSRA team partners with local under-served communities around the world to improve the access to and quality of knowledge based on education, amplify and augment learning programs where they exist, and create new opportunities for e-learning where traditional education systems are lacking or non-existent.

FULL Paper PDF file:




A. Menon and B. Omman




Detection and Recognition of Multiple License Plate From Still Images

Publish in

2018 International Conference on Circuits and Systems in Digital Enterprise Technology (ICCSDET), Kottayam, India, 2018, pp. 1-5,



PDF reference and original file: Click here



+ posts

Somayeh Nosrati was born in 1982 in Tehran. She holds a Master's degree in artificial intelligence from Khatam University of Tehran.

Website | + posts

Professor Siavosh Kaviani was born in 1961 in Tehran. He had a professorship. He holds a Ph.D. in Software Engineering from the QL University of Software Development Methodology and an honorary Ph.D. from the University of Chelsea.

Website | + posts

Nasim Gazerani was born in 1983 in Arak. She holds a Master's degree in Software Engineering from UM University of Malaysia.