article · IEEE Access
Most license plate recognition systems perform well only under ideal conditions with well-annotated data, whereas monitoring systems frequently capture low-resolution images in naturalistic environments. To address this limitation, an end-to-end deep learning framework combines single-stage character segmentation and recognition with super-resolution generative adversarial networks. The architecture incorporates architectural modifications, including adjustments to the number of layers, activation functions, and the addition of total variation loss regularisation. This approach converts low-resolution inputs, including small 72 by 72 pixel images, into realistic high-resolution versions. Evaluations across multiple datasets using visual analysis, peak signal-to-noise ratio, structural similarity index measure, and optical character recognition show that the method reliably enhances image quality and improves license plate recognition accuracy over existing systems.
Real-world surveillance footage frequently suffers from low resolution, preventing automated systems from identifying vehicle licence plates accurately. Reconstructing clearer, high-resolution imagery from degraded footage enhances recognition accuracy in unconstrained environments. This capability helps organisations improve the reliability of automated vehicle monitoring and media forensic analysis without requiring immediate upgrades to camera hardware.
The method is applicable to automated vehicle monitoring, traffic enforcement, and media forensic software where camera footage is low quality. Potential end users include municipal surveillance operators, forensic examiners, and security technology vendors. Having been developed and evaluated across several datasets, the framework sits at an applied and tested stage, serving as an algorithmic enhancement for existing image recognition pipelines.
AI-generated from the published abstract. Always read the original work before citing.
Although the majority of existing License Plate (LP) recognition techniques have significant improvements in accuracy, they are still limited to ideal situations in which training data is correctly annotated with restricted scenarios. Moreover, images or videos are frequently used in monitoring systems that have Low Resolution (LR) quality. In this work, the problem of LP detection in digital images is addressed in the images of a naturalistic environment. Single-stage character segmentation and recognition are combined with adversarial Super-Resolution (SR) approaches to improve the quality of the LP by processing the LR images into High-Resolution (HR) images. This work proposes effective changes to the SRGAN network regarding the number of layers, an activation function, and the appropriate loss regularization using Total Variation (TV) loss. The main paper contribution can be summarized into presenting an end-to-end deep learning framework based on generative adversarial networks (GAN), which is able to generate realistic super-resolution images. Also, proposed adding a TV regularization to the loss function to help the model enhance the resolution of images. The proposed SRGAN can handle tiny <inline-formula> <tex-math notation="LaTeX">$72\times 72$ </tex-math></inline-formula> images of LPs. The paper explores how SRGAN performed over different datasets from many aspects, such as visual analysis, PSNR, SSIM, and Optical Character Recognition (OCR). The experiments demonstrate that the suggested SRGAN can generate high-resolution images that improve the accuracy of the license plate recognition stage compared to other systems.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1109/access.2022.3157714
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
