License Plate Image Reconstruction Based on Generative Adversarial Networks

Mianfen Lin,Liangxin Liu,Jiahui Pan,Jingcong Li,Fei Wang

doi:10.3390/rs13153018

Mianfen Lin, Liangxin Liu + Show 3 more

Open Access

https://doi.org/10.3390/rs13153018

Copy DOI

Abstract

License plate image reconstruction plays an important role in Intelligent Transportation Systems. In this paper, a super-resolution image reconstruction method based on Generative Adversarial Networks (GAN) is proposed. The proposed method mainly consists of four parts: (1) pretreatment for the input image; (2) image features extraction using residual dense network; (3) introduction of progressive sampling, which can provide larger receptive field and more information details; (4) discriminator based on markovian discriminator (PatchGAN) can make a more accurate judgment, which guides the generator to reconstruct images with higher quality and details. Regarding the Chinese City Parking Dataset (CCPD) dataset, compared with the current better algorithm, the experiment results prove that our model has a higher peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) and less reconstruction time, which verifies the feasibility of our approach.

Highlights

License plate image reconstruction plays an important role in Intelligent Transportation Systems
Chinese City Parking Dataset (CCPD) dataset, compared with the current better algorithm, the experiment results prove that our model has a higher peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) and less reconstruction time, which verifies the feasibility of our approach
(2) Simultaneously, for the special image of a license plate, we focus on advanced features such as edge and letter direction in order to minimize the mixture of VGG loss and mean square error (MSE) as the loss function parameter of our algorithm

Summary

Introduction

License plate image reconstruction plays an important role in Intelligent Transportation Systems. Discriminator based on markovian discriminator (PatchGAN) can make a more accurate judgment, which guides the generator to reconstruct images with higher quality and details. License plate recognition is a significant component of an intelligent transportation system It is widely used in such fields as vehicle positioning, vehicle identification at highway toll stations and speed measurement. Taking full advantage of deep learning technology to repair and reconstruct blurry license plate images, it is convenient to perform subsequent positioning or other operations on the captured vehicles. This is realistic for maintaining social traffic safety and realizing traffic automation management. In the field of medical images, doctors need the details of images to make accurate judgments

Objectives

Results

Discussion

Conclusion