License Plate Image Analysis Empowered by Generative Adversarial Neural Networks (GANs)

License plate recognition is currently used in various situations, such as parking lot vehicle management and the tracking of wanted vehicles. These applications require recognition from all possible camera angles. Therefore, we propose a method for license plate recognition in images captured from various camera angles. First, the license plate area is detected from the input image using the YOLOv5 object detection method. The subsequent process can be performed in two manners: by rotating the license plate to the front using projective transformation and trapezoidal correction, or by using graph matching. The first method performs Hough transforms or linear approximations as preprocessing, and then calculates the rectangle surrounding the license plate from the YOLOv5 output. After performing trapezoidal correction, the maximally stable extremal regions (MSER) are used to detect character candidates, and characters are recognized using 3D rotated character recognition. The second method performs character candidate detection and recognition without trapezoidal correction, followed by graph matching. These methods are versatile because they do not depend on the layout of license plates, which varies among countries and regions. Two types of datasets were used: a set of images containing Japanese license plates collected by ourselves, and a set of publicly available images containing Taiwanese license plates. Two recognition rates were evaluated: the license plate character recognition rate and license plate recognition rate in which all characters in one plate are recognized. The license plate character recognition rates using graph matching were 91.9% for the Japanese license plates and 89.3% for the Taiwanese license plates. The license plate recognition rates were 84.0% and 84.5%, respectively.

Background: Most previous studies in automatic license plate recognition (ALPR) focused on recognizing license plate (LP) in constrained environment where cameras are installed in front of LPs and other conditions such as lighting, weather, and image quality are satisfied. Besides, recent studies on ALPR in Vietnam have conducted in small datasets and have not covered various cases of Vietnamese LPs.Aim: To develop a model for ALPR that is effective in unconstrained environment in Vietnam.Method: We propose two improvements: We apply the idea of the key-point detection problem for LP detection part, and use a segmentation free approach based on encoder decoder network for the LP optical character recognition (OCR) part. We train and evaluate models in a large dataset collected from unconstrained environment.Results: Our results show improvements in LP detection accuracy with mean IOU mIOU = 95.01% and precision P <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">75</inf> = 99, 5%. The accuracy in LP OCR was up to Acc <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">seq</inf> = 99.28% at sequence level and Acc <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">char</inf> = 99.7% at character level.Conclusion: We provide a large dataset of Vietnamese LP images that can be effectively used to evaluate ALPR systems in Vietnam, and proposes improvement techniques to tackle problems of ALPR in unconstrained environment in Vietnam.

Traditional license plate recognition technology mostly uses traditional image processing methods to find out the characteristics of the license plate, and then crop and recognize the characters. The process needs to be modified due to the different environments, scenes and conditions. In recent years, many studies have implemented license plate and character recognition by using deep learning algorithms. Although it has a good recognition accuracy, the calculation speed still cannot reach the level of real-time recognition. This research proposes a real-time license plate recognition system based on YOLOv3, which uses deep learning model to realize the vehicle license plate recognition, lane identification and vehicle trajectory tracking. In this study, a web-based platform is established to present the result of license plate recognition and trajectory, and the streaming roadside video in the campus. In the platform, license plates of driving vehicles can be identified in real-time, and the user can search and track specific vehicle intuitively. In the experiment, the average accuracy of the system performs 84.3% in real-time license plate recognition, and 100% in lane identification. The system can process in 40 FPS, which can meet the level of real-time system. In the future, the system can cooperate with traffic access control in campus or community to improve the efficiency of traffic control.

Automatic License Plate Recognition (ALPR) is a very widely used system in applications such as parking management, theft detection, traffic control and management etc. Most of the existing ALPR systems fail to showcase acceptable performance on real time images/video scenes. This work proposes and demonstrates implementation of a deep learning-based approach to locate license plates of four wheeler vehicles thereby enabling optical character recognition (OCR) to recognize the characters and numbers on the located plates in real time. The proposed system is decomposed into three sub-blocks viz. Vehicle image/video acquisition, License plate localization and OCR. A simple setup using a reasonable resolution webcam has been designed to capture images/videos of vehicles at some entry point. We propose to utilize Single Shot Detector (SSD) based Mobilenet V1 architecture to localize the license plates. The hyper parameters of this architecture are selected with rigorous experimentation so as to avoid over-fitting. We have compared performance of two OCRs viz. Tesseract OCR, Easy OCR and found the superiority of Easy OCR since it utilizes deep learning approach for character recognition. NVIDIA Jetson Nano and Raspberry Pi 3B hardware platforms have been used to implement the entire system. The parameters of these three sub-blocks have been optimized to yield real time performance of ALPR with acceptable accuracy. The proposed and implemented system on Jetson Nano allows processing of videos for ALPR having accuracy more than 95%.

The city's efforts to improve traffic make a big step forward when license plates can be read. It explains how an intelligent transportation system should work and what steps should be taken for it to be successful. Because the number of cars is growing so quickly, ANPR is a must-have for managing traffic control. Automatic number plate recognition (ANPR) is mostly used to keep an eye on traffic and keep people safe. Number plate recognition uses image processing and the most recent advances in technology to automatically read the characters on a vehicle's license plate. In the last few years, there have been a number of technological advances in the field of studying how to read license plates. Image processing techniques like OCR make it possible for traffic surveillance to solve a wide range of problems that come up during criminal investigations, the collection of tolls, the monitoring of traffic, the regulation of speed, and the management of parking, among other things. If you want to control traffic and keep an eye on a large number of people in a transportation system, you need an ANPR system. Image processing techniques and the collection of photos of vehicles for use in the dataset have made it possible to keep an eye on traffic on a large scale. Automatic License Plate Reader (ANPR) is a way to take the steps needed to run a good intelligent transportation network. Traffic control has become an absolute must because the number of cars has grown so quickly. The main goal of ANPR is to keep an eye on and record traffic for defense purposes. To read the text on license plates, number plate recognition uses image processing, optical character recognition (OCR), and edge detection technology. The model is made up of three different parts called, respectively, the module for car detection, the module for license plate segmentation, and the module for recognition. Image processing made us more determined to stop a wide range of illegal activities, such as armed robberies of cars, breaking traffic rules, and the way law enforcement is handled. This review study looked at all of the different designs for reading license plates that have been put into use so far.

This paper presents an efficient method for license plate detection and recognition that will aid in the development of an intelligent transport system. To significantly accelerate license plate localization a simple image pre-processing method is first used without reducing detection performance compared with that obtained using original image. Line density filter approach is used to extract the license plate regions, there by significantly reducing area to be analysed for license plate localization. In license plate recognition, the characters are recognized from these detected license plates by using optical character recognition and the owner details are displayed. The aim of this paper is to serve the purpose of accurately localizing vehicle license plates from complex scenes in real time that would help motor vehicle department in various ways.

Automatic License Plate Recognition (ALPR) has been the focus of many\nresearches in the past years. In general, ALPR is divided into the following\nproblems: detection of on-track vehicles, license plates detection, segmention\nof license plate characters and optical character recognition (OCR). Even\nthough commercial solutions are available for controlled acquisition\nconditions, e.g., the entrance of a parking lot, ALPR is still an open problem\nwhen dealing with data acquired from uncontrolled environments, such as roads\nand highways when relying only on imaging sensors. Due to the multiple\norientations and scales of the license plates captured by the camera, a very\nchallenging task of the ALPR is the License Plate Character Segmentation (LPCS)\nstep, which effectiveness is required to be (near) optimal to achieve a high\nrecognition rate by the OCR. To tackle the LPCS problem, this work proposes a\nnovel benchmark composed of a dataset designed to focus specifically on the\ncharacter segmentation step of the ALPR within an evaluation protocol.\nFurthermore, we propose the Jaccard-Centroid coefficient, a new evaluation\nmeasure more suitable than the Jaccard coefficient regarding the location of\nthe bounding box within the ground-truth annotation. The dataset is composed of\n2,000 Brazilian license plates consisting of 14,000 alphanumeric symbols and\ntheir corresponding bounding box annotations. We also present a new\nstraightforward approach to perform LPCS efficiently. Finally, we provide an\nexperimental evaluation for the dataset based on four LPCS approaches and\ndemonstrate the importance of character segmentation for achieving an accurate\nOCR.\n

Development of high quality license plate recognition system is a challenging task, not fully solved nowadays. License plate recognition process consists of the following steps: license plate detection, individual characters segmentation and recognition. This paper contains methods, connected with license plate allocation, segmentation and characters recognition. The noise on the plate and its angular inclination are main problems raised during developing such systems. In this article a new method of license plate recognition is presented. The proposed method includes preliminary image filtering, connected component method for segmentation and hierarchical temporal memory model for recognition. Image prefiltering improves the efficiency of subsequent binarization. Generally, license plate segmentation is provided by the histogram method, with different angles of inclination of the registration plate. As a result the rotation of the plate reduces the image quality. The connected component method eliminates rotation from this process, and provides no loss of image quality. Separate symbols can be represented under a small angle after such segmentation, which could complicate their identification. However, the application of the hierarchical temporal memory model for character recognition, previously trained at the sloping characters images, gives positive results. The proposed algorithms can also be used for distorted text segmentation and recognition.

The paper is devoted to the problem of automatic detection and recognition of license plates, the solution of which has many potential applications, from security to traffic management. The purpose of this work was to compare the methods of finding and recognizing car number plates, based on the application of deep learning algorithms, which takes into account different regional standards of car number plates, video quality, different speeds of vehicles, the location of the camera in relation to the vehicle license plate, defects of the car number plate (pollution , deformation), as well as changes in external lighting conditions. The advantages and disadvantages of localization and segmentation of car number plates on cars using image binarization, Viola–Jones and Harr methods are given. It was determined that adaptive approaches are better due to the possibility of compensating the impact of obstacles on different areas of the image, for example, the distribution of shadows due to the heterogeneity of illumination. It was determined that many methods in real algorithms rely directly or indirectly on the presence of number limits. Even if the limits are not used when the number is determined, they have the possibility to be used for further analysis. The methods of templates, image histograms, and contour analysis were compared to identify familiar features in the image (segmentation). It is shown that an effective approach for recognition of car license plates can be based on the application of the methods of Viola-Jones, Harr, the analysis of brightness histograms and the SVM method. Formulated conclusions on the effectiveness of the implementation of each of the procedures were confirmed as a result of conducting experiments with the developed software in the python 3 language using the cv2 computer vision library. The described approach makes it possible to obtain a fairly high recognition accuracy at different angles of rotation of the license plate relative to the camera. Keywords: automatic recognition, license plates, localization, normalization, segmentation, character recognition.

Automatic License plate detection and recognition system is special form of optical character recognition and has been an active research domain in image processing field. However, the accuracy is varied due to different styles of number plates endorsed. Besides, the characters on Malaysia license plate are in one or two lines. Thus, the proposed license plate recognition (LPR) technique of this research is able to achieve the best recognition performance based on Malaysia license plate vehicle registration number characters. This paper presents a study of applying the template matching method for character image recognition. The database of characters and license plate image has been created by collecting images from various type of car. The initial pre-processing involves image enhancement, binarization, filtering and segmenting of license plate. There are 100 license plates that contain 693 characters have been tested, and the result shown that 92.78 % of all characters is correctly recognized. Thus, template matching can be classified as one of the promising algorithm for recognizing Malaysia license plate.

On the streets of Taiwan, many roadside tollers are often seen riding motorcycles in one hand, and the other hand holding mobile devices to issue payment notices for cars and motorcycles parked on the roadsides. The work of roadside tollers is very dangerous. First, they must first park their motorcycles next to the roadside cars and motorcycles. They then use their eyes to confirm the license plate number, enter the license plate number into the mobile device, and finally place the bill on the car’s windows or attach the bill to the motorcycles. Our idea is to implement an automated license plate recognition system in mobile devices to increase the efficiency of roadside tollers and reduce their time on the road. Recently, license plate recognition systems have been widely used in various aspects of life, such as parking lot toll systems, access management systems, and traffic management systems. However, existing license plate recognition systems must have good recognition rates under a number of constraints, such as fixed angles and fixed light sources. Moreover, due to the insufficient computing resources of the general mobile device, the application cannot have a good recognition rate in the complex environment or skewed angle in the license plate recognition. Therefore, this paper proposes a lightweight and high-performance multi-angle license plate character recognition model, which reduces the complexity and computational complexity of traditional license plate recognition. This paper also collects a large number of license plate images from different environments, angles and sizes as training data. Finally, we propose an optimized deep learning model to identify the characters on license plates. The experimental results show that the proposed model can recognize the license plate with a tilt of 0~60 degrees, and the overall recall rate is 84.5%. Compared with Tiny-YOLOv2, the computation of the proposed model is reduced by 61% with a little penalty of recall.

CIS Multilingual License Plate Detection and Recognition Based on Convolutional and Transformer Neural Networks

Almost every vehicle has currently installed black box since the stored images by black box can be used to investigate the exact cause of the accident. One of the most important aspects in an accident investigation is the license plate detection and recognition as the license plate has information about the driver and car. This paper presents a novel algorithm for license plate detection and recognition using black box image. The proposed license plate recognition system is divided into three stages: license plate detection, individual number and character extraction, and number and character recognition. The Gaussian blur filter is used to remove noise in the image and then we detect the license plate edge using modified Canny algorithm. Second, we determine license plate candidate image using morphology and support vector machine. Finally, we recognize the numbers and characters using k-nearest neighbor classifier. The experimental study results indicate that the license plate detection and recognition algorithm has been successfully implemented.

Efficient and accurate license plate recognition (LPR) in unconstrained environments remains a critical challenge, particularly when confronted with skewed imaging angles and the limited computational capabilities of edge devices. In this study, we propose a high-performance, FPGA-based license plate alignment and recognition (LPAR) system to address these issues. Our LPAR system integrates lightweight deep learning models, including YOLOv4-tiny for license plate detection, a refined convolutional pose machine (CPM) for pose estimation and alignment, and a modified LPRNet for character recognition. By restructuring the pose estimation and alignment architectures to enhance the geometric correction of license plates and adding channel and spatial attention mechanisms to LPRNet for better character recognition, the proposed LPAR system improves recognition accuracy from 88.33% to 95.00%. The complete pipeline achieved a processing speed of 2.00 frames per second (FPS) on a resource-constrained FPGA platform, demonstrating its practical viability for real-time deployment in edge computing scenarios.

Recent years have seen significant developments in license plate recognition through the integration of deep learning techniques and the increasing availability of training data. Nevertheless, reconstructing license plates from low-resolution surveillance footage remains a challenge. To address this issue, we propose an attention-based super-resolution approach that incorporates sub-pixel convolution layers and an Optical Character Recognition (OCR)-based loss function. We trained the proposed architecture using synthetic images created by applying heavy Gaussian noise followed by bicubic downsampling to high-resolution license plate images. Our results show that the proposed approach for reconstructing these low-resolution images substantially outperforms existing methods in both quantitative and qualitative measures. Our source code is publicly available at https://github.com/valfride/lpr-rsr-ext/.