Next-Generation License Plate Detection and Recognition System using YOLOv8
📝 Abstract
In the evolving landscape of traffic management and vehicle surveillance, efficient license plate detection and recognition are indispensable. Historically, many methodologies have tackled this challenge, but consistent real-time accuracy, especially in diverse environments, remains elusive. This study examines the performance of YOLOv8 variants on License Plate Recognition (LPR) and Character Recognition tasks, crucial for advancing Intelligent Transportation Systems. Two distinct datasets were employed for training and evaluation, yielding notable findings. The YOLOv8 Nano variant demonstrated a precision of 0.964 and mAP50 of 0.918 on the LPR task, while the YOLOv8 Small variant exhibited a precision of 0.92 and mAP50 of 0.91 on the Character Recognition task. A custom method for character sequencing was introduced, effectively sequencing the detected characters based on their x-axis positions. An optimized pipeline, utilizing YOLOv8 Nano for LPR and YOLOv8 Small for Character Recognition, is proposed. This configuration not only maintains computational efficiency but also ensures high accuracy, establishing a robust foundation for future real-world deployments on edge devices within Intelligent Transportation Systems. This effort marks a significant stride towards the development of smarter and more efficient urban infrastructures.
💡 Analysis
In the evolving landscape of traffic management and vehicle surveillance, efficient license plate detection and recognition are indispensable. Historically, many methodologies have tackled this challenge, but consistent real-time accuracy, especially in diverse environments, remains elusive. This study examines the performance of YOLOv8 variants on License Plate Recognition (LPR) and Character Recognition tasks, crucial for advancing Intelligent Transportation Systems. Two distinct datasets were employed for training and evaluation, yielding notable findings. The YOLOv8 Nano variant demonstrated a precision of 0.964 and mAP50 of 0.918 on the LPR task, while the YOLOv8 Small variant exhibited a precision of 0.92 and mAP50 of 0.91 on the Character Recognition task. A custom method for character sequencing was introduced, effectively sequencing the detected characters based on their x-axis positions. An optimized pipeline, utilizing YOLOv8 Nano for LPR and YOLOv8 Small for Character Recognition, is proposed. This configuration not only maintains computational efficiency but also ensures high accuracy, establishing a robust foundation for future real-world deployments on edge devices within Intelligent Transportation Systems. This effort marks a significant stride towards the development of smarter and more efficient urban infrastructures.
📄 Content
Next-Generation License Plate Detection and Recognition System using YOLOv8 Arslan Amin School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), Islamabad, 44000, Pakistan Email:aamin.mscs19seecs@seecs.edu.pk Rafia Mumtaz School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), Islamabad, 44000, Pakistan Email:rafia.mumtaz@seecs.edu.pk Muhammad Jawad Bashir School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), Islamabad, 44000, Pakistan Email:mbashir.msds20seecs@seecs.edu.pk Syed Mohammad Hassan Zaidi Ghulam Ishaq Khan Institute of Engineering Sciences and Technology (GIKI), Topi, District, Swabi, Khyber Pakhtunkhwa 23460, Pakistan, Email:prorector acad@giki.edu.pk Abstract—In the evolving landscape of traffic management and vehicle surveillance, efficient license plate detection and recognition are indispensable. Historically, many methodologies have tackled this challenge, but consistent real-time accuracy, especially in diverse environments, remains elusive. This study examines the performance of YOLOv8 variants on License Plate Recognition (LPR) and Character Recognition tasks, crucial for advancing Intelligent Transportation Systems. Two distinct datasets were employed for training and evaluation, yielding notable findings. The YOLOv8 Nano variant demonstrated a precision of 0.964 and mAP50 of 0.918 on the LPR task, while the YOLOv8 Small variant exhibited a precision of 0.92 and mAP50 of 0.91 on the Character Recognition task. A custom method for character sequencing was introduced, effectively sequencing the detected characters based on their x-axis positions. An optimized pipeline, utilizing YOLOv8 Nano for LPR and YOLOv8 Small for Character Recognition, is proposed. This configuration not only maintains computational efficiency but also ensures high accuracy, establishing a robust foundation for future real-world deployments on edge devices within Intelligent Transportation Systems. This effort marks a significant stride towards the development of smarter and more efficient urban infrastructures. Index Terms—License Plate Detection, Computer Vision, YOLOv8, Object Detection, Vehicle Surveillance, Character Recognition In the intricate web of modern urban infrastructure, the ca- pability to identify vehicles swiftly and accurately has emerged as a cornerstone for numerous applications. From traffic man- agement and congestion control to security surveillance and automated toll collection, the importance of efficient license plate detection and recognition has grown exponentially [1]. Traditional methods of license plate detection, while com- mendable in their efforts, often grapple with challenges such as variability in environmental conditions, diverse license plate designs across regions, and the need for real-time processing. The dynamic nature of these challenges has fueled a relentless pursuit of more adaptive and robust solutions. The field of License Plate Recognition (LPR) has witnessed a significant evolution, pivoting from the foundational tasks of detection, extraction, and recognition of license plates to embracing more nuanced challenges such as deblurring, denoising, and geometric transformations to bolster recog- nition accuracy [2]. Initially, the exploration in this domain was channeled through employing neural networks to tackle individual facets of the problem. However, a paradigm shift occurred as subsequent efforts aimed at orchestrating neural networks to oversee the entire process, with proposals extend- ing to utilizing up to five networks for discerning the presence of a license plate within an image. Various neural network architectures such as Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), and Recurrent Neu- ral Networks (RNNs) have been utilized in License Plate Recognition (LPR). CNNs excel in character detection and recognition, GANs can generate synthetic training data, and RNNs may aid in recognizing character sequences on license plates. A notable trajectory in this field has been the strategic modification of the YOLO [3] network to accelerate processing speed, reflecting a concerted effort to match the real-time exigencies of surveillance and traffic monitoring applications. Deep learning, particularly in the domain of CNNs, has shown immense promise in revolutionizing object detection tasks, including license plate recognition. Among the various architectures and models that have gained prominence, the YOLO series stands out due to its ability to detect objects in real-time with remarkable accuracy. Moreover, the ingenuity embedded in methods like N-YOLO [4], which deviates from traditional image size reduction to employing fixed-size image patches, and innovative approaches like CornerNet [5] and HoughNet [6], that shift the gaze from top-bottom to b
This content is AI-processed based on ArXiv data.