A Robust Deep Learning Approach for Automatic Iranian Vehicle License Plate Detection and Recognition for Surveillance Systems

Abstract

The process of detecting vehicles' license plates, along with recognizing the characters inside them, has always been a challenging issue due to various conditions. These conditions include different weather and illumination, inevitable data acquisition noises, and some other challenging scenarios like the demand for real-time performance in state-of-the-art Intelligent Transportation Systems (ITS) applications. This paper proposes a method for vehicle License Plates Detection (LPD) and Character Recognition (CR) as a unified application that presents significant accuracy and real-time performance. The mentioned system is designed for Iranian vehicle license plates, which have the characteristics of different resolution and layouts, scarce digits/characters, various background colors, and different font sizes. In this regard, the system uses a separate fine-tuned You Only Look Once (YOLO) version 3 platform for each of the mentioned phases and extracts Persian characters from input images in two automatic steps. For training and testing stages, a wide range of vehicle images in different challenging and straightforward conditions have been collected from practical systems installed as surveillance applications. Experimental results show an end-to-end accuracy of 95.05% on 5719 images. The test data included both color and grayscale images containing the vehicles with different distances and shooting angles with various brightness and resolution. Additionally, the system can perform the LPD and CR tasks in an average of 119.73 milliseconds for real life data, which illustrates a real-time performance for the system and usable applicability. The system is fully automated, and no pre-processing, calibration or configuration procedures are needed.

License Plate Detection Using Convolutional Neural Network–Back to the Basic With Design of Experiments

“…[20] further enhance datasets by synthesising LPs to overcome small dataset size and train a custom CNN model ported to Fast-YOLO to perform ALPR. [21] utilised YOLOv3 for both LP detection and recognition stages with 95-97% accuracy. Overall, the YOLO-based algorithm is very promising to be repurposed for LP detection.…”

Iranis: A Large-scale Dataset of Iranian Vehicles License Plate Characters

“…The authors of this paper have contributed to many computer vision and machine learning projects and proposed various approaches in the field of ITS. Some of these approaches include vehicle count using video processing [16], deep learning-based vehicle detection [17], vehicle speed measurement [18][19], license plate localization [8,20], and Farsi character recognition [8]. Accordingly, we claim that we have felt the essence of reliable data for the development of domestic robust applications for Fig.…”

“…To obtain characters from license plates, we have proposed a two-stage deep learning approach that localizes plates in images, and then segments and extracts the inside texts. Further information about these stages can be found in our recently published work [8]. Table I illustrates the characteristics of the dataset in detail.…”

Towards Automatic License Plate Detection

“…Many of them perform well in constrained environments, such as a single license plate in an input image with a simple background, fixed illumination, and a slightly distorted/blurred license plate [ 4 ]. Recent state-of-the-art techniques, such as [ 4 , 5 , 6 ], put less limits on object/license plate detection at the expense of higher computing complexity. Moreover, extracting license plates from complicated scenes remains a significant challenge for these methods.…”