计算机科学 ›› 2026, Vol. 53 ›› Issue (6A): 250600070-5.doi: 10.11896/jsjkx.250600070
黄海新, 侯广帅, 何添禹
HUANG Haixin, HOU Guangshuai, HE Tianyu
摘要： 在智慧交通系统中,由于监控摄像头捕捉的车牌图像存在光照影响、运动模糊、获取的车牌图像分辨率低等问题,超分辨率重构在车牌图像的研究受到广泛关注。现有超分辨率方法往往存在重建图像存在伪影、高频细节丢失导致图像细节边缘模糊等问题,因此提出了SeguGAN框架。首先,由预训练视觉模型CLIP提取车牌语义信息,在判别器中添加语义感知模块(SeMSCA)融合语义信息与图像特征,提升判别器能力;其次,在生成器中引入门控机制融合两个不同分支(RRDB,CAMConv)提取的图像特征,提升图像重建质量。最后,使用动态Tanh(DyT)代替传统层归一化,简化模型,提升模型能力。为验证Segu-GAN的有效性,在公开数据集CCPD2019和CCPD2020上进行实验,实验中SeguGAN达到33.20 dB PSNR与0.906 SSIM,较主流的ESRGAN,SRGAN,ECBSR,RCAN,SwinIR模型PSNR平均提升5.9%,SSIM平均提升1.9%。该结果证实所提方法能够优化车牌图像的超分辨率重建效果。
中图分类号:
| [1] ZHANG X,ZENG H,ZHANG L.Edge-oriented convolutionblock for real-time super resolution on mobile devices[C]//Proceedings of the 29th ACM International Conference on Multimedia.2021:4034-4043. [2] ALHALAWANI S,BENJDIRA B,AMMAR A,et al.DiffPlate:A Diffusion Model for Super-Resolution of License Plate Images [J].Electronics,2024,13(13):2670. [3] GOODFELLOW I J,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets [C]//Proceedings of the 28th International Conference on Neural Information Processing Systems.2014:2672-2680. [4] LEDIG C,THEIS L,HUSZÁR F,et al.Photo-realistic singleimage super-resolution using a generative adversarial network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern recognition.2017:4681-4690. [5] WANG X,YU K,WU S,et al.Esrgan:Enhanced super-resolution generative adversarial networks[C]//Proceedings of the European Conference on Computer Vision(ECCV).2018. [6] KUZNEDELEV D,STARTSEV V,SHLENSKII D,et al.Does Diffusion Beat GAN in Image Super Resolution? [J].arXiv:2405,17261,2024. [7] DONG C,LOY C C,HE K,et al.Learning a deep convolutional network for image super-resolution[C]//European Conference on Computer Vision.2014:184-199. [8] ZHANG Y,LI K,LI K,et al.Image super-resolution using very deep residual channel attention networks[C]//Proceedings of the European Conference on Computer Vision(ECCV).2018:286-301. [9] ZHANG W,LIU Y,DONG C,et al.Ranksrgan:Generative adversarial networks with ranker for image super-resolution[C]//Proceedings of the IEEE/CVF Inernational Conference on Computer Vision.2019:3096-3105. [10] LIANG J,CAO J,SUN G,et al.Swinir:Image restoration using swin transformer[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2021:1833-1844. [11] LIU X,LIU J,TANG J,et al.CATANet:Efficient Content-Aware Token Aggregation for Lightweight Image Super-Resolution[C]//Proceedings of the Computer Vision and Pattern Re-cognition Conference.2025:17902-17912. [12] LI X,WANG Z,ZOU Y,et al.Difiisr:A diffusion model with gradient guidance for infrared image super-resolution[C]//Proceedings of the Computer Vision and Pattern Recognition Conference.2025:7534-7544. [13] YUE Z,LIAO K,LOY C C.Arbitrary-steps image super-resolution via diffusion inversion[C]//Proceedings of the Computer Vision and Pattern Recognition Conference.2025:23153-23163. [14] LEE S,KIM J-H,HEO J-P.Super-resolution of license plate images via character-based perceptual loss[C]//2020 IEEE International Conference on Big Data and Smart Computing(BigComp).2020:560-563. [15] XU S J,DENGB W,SHI Y,et al.A Super-resolution Network for License Plate Images with an Encoder-decoder Structure [J].Journal of Xi'an Jiaotong University,2022,56(10):101-110. [16] PAN Y,TANG J,TJAHJADI T.LPSRGAN:Generative adversarial networks for super-resolution of license plate image [J].Neurocomputing,2024,580:127426. [17] RADFORD A,KIM J W,HALLACY C,et al.Learning transferable visual models from natural language supervision[C]//International Conference on Machine Learning.2021:8748-8763. [18] LI B,LI X,ZHU H,et al.Sed:Semantic-aware discriminator for image super-resolution[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2024:25784-25795. [19] ZHU J,CHEN X,HE K,et al.Transformers without normalization[C]//Proceedings of the Computer Vision and Pattern Recognition Conference.2025:14901-14911. [20] JANG E,GU S,POOLE B.Categorical reparameterization with gumbel-softmax [J].arXiv:1611,01144,2016. [21] LV X,WANG B,ZOU J Y,et al.License Plate Image Clarity Enhancement Based on Improved Attention Mechanism [J].Radio Engineering,2021,51(10):1169-1175. [22] WANG X,XIE L,DONG C,et al.Real-esrgan:Training real-world blind super-resolution with pure synthetic data[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2021:1905-1914. |
|
||
首页