Abstract:Abstract Non-standard license plates are a part of current traffic trends in Pakistan. Private number plates should be recognized and, monitored for several purposes including security as well as a well-developed traffic system. There is a challenging task for the authorities to recognize and trace the locations for the certain number plate vehicle. In a developing country like Pakistan, it is tough to have higher constraints on the efficiency of any license plate identification and recognition algorithm. Character recognition efficiency should be a route map for the achievement of the desired results within the specified constraints. The main goal of this study is to devise a robust detection and recognition mechanism for non-standard, transitional vehicle license plates generally found in developing countries. Improvement in the character recognition efficiency of drawn and printed plates in different styles and fonts using single using multiple state-of-the-art technologies including machine-learning (ML) models. For the mentioned study, 53-layer deep convolutional neural network (CNN) architecture based on the latest variant of object detection algorithm-You Only Look Once (YOLOv3) is employed. The proposed approach can learn the rich feature representations from the data of diversified license plates. The input image is first pre-processed for quality improvement, followed by dividing it into suitable-sized grid cells to find the correct location of the license plate. For training the CNN, license plate characters are segmented. Lastly, the results are post-processed and the accuracy of the proposed model is determined through standard benchmarks. The proposed method is successfully tested on a large image dataset consisting of eight different types of license plates from different provinces in Pakistan. The proposed system is expected to play an important role in implementing vehicle tracking, payment for parking fees, detection of vehicle over-speed limits, reducing road accidents, and identification of unauthorized vehicles. The outcome shows that the proposed approach achieves a plate detection accuracy of 97.82% and the character recognition accuracy of 96%.

License Plate Detection for Multi-National Vehicles: an Illumination Invariant Approach in Multi-Lane Environment.

License Plate Detection for Multi-National Vehicles – a Generalized Approach

A Fast Vehicle License Plate Locating Algorithm Under Complex Environments

A Multi-Style License Plate Verification Method Based on Character Classification under Illumination Variances

Multinational vehicle license plate detection in complex backgrounds.

A Location Method for Multi-style License Plates Based on Character Classification

An Adaptive Approach for Multi-National Vehicle License Plate Recognition Using Multi-Level Deep Features and Foreground Polarity Detection Model

Chinese License Plate Localization in Multi-Lane with Complex Background Based on Concomitant Colors

License plate identification and recognition in a non-standard environment using neural pattern matching

A Robust and Efficient Approach to License Plate Detection

Multiple Licence Plate Detection for Chinese Vehicles in Dense Traffic Scenarios

A New CNN-Based Method for Multi-Directional Car License Plate Detection

Performance Evaluation of Local Image Features for Multinational Vehicle License Plate Verification

Scale-Invariant Multidirectional License Plate Detection with the Network Combining Indirect and Direct Branches

Towards Automatic License Plate Detection

Multi-Oriented and Scale-Invariant License Plate Detection Based on Convolutional Neural Networks

Recognizing License Plates in Real-Time

A Novel and Robust Algorithm for License Plate Location Using Perceptual Salient Features

A Robust License Plate Detection and Character Recognition Algorithm Based on a Combined Feature Extraction Model and BPNN

An End to End Recognition for License Plates Using Convolutional Neural Networks

License Plate Localization in Unconstrained Scenes Using a Two-Stage CNN-RNN