Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging
The conventional denoising method in Rayleigh imaging in a general sense requires an additional hardware investment and the use of the underlying physics. This work demonstrates an alternative image denoising reconstruction model based on unsupervised learning that aims to remove Mie scattering and...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-08-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/15/15/5747 |
| _version_ | 1797442190873460736 |
|---|---|
| author | Minnan Cai Hua Jin Beichen Lin Wenjiang Xu Yancheng You |
| author_facet | Minnan Cai Hua Jin Beichen Lin Wenjiang Xu Yancheng You |
| author_sort | Minnan Cai |
| collection | DOAJ |
| description | The conventional denoising method in Rayleigh imaging in a general sense requires an additional hardware investment and the use of the underlying physics. This work demonstrates an alternative image denoising reconstruction model based on unsupervised learning that aims to remove Mie scattering and shot noise interference from two-dimensional (2D) Rayleigh images. The model has two generators and two discriminators whose parameters can be trained with either feature-paired or feature-unpaired data independently. The proposed network was extensively evaluated with a qualitative examination and quantitative metrics, such as PSNR, ER, and SSIM. The results demonstrate that the feature-paired training network exhibits a better performance compared with several other networks reported in the literature. Moreover, when the flame features are not paired, the feature-unpaired training network still yields a good agreement with ground truth data. Specific indicators of the quantitative evaluation show a promising denoising ability with a peak signal-to-noise ratio of ~37 dB, an overall reconstruction error of ~1%, and a structure similarity index of ~0.985. Additionally, the pre-trained unsupervised model based on unpaired training can be generalized to denoise Rayleigh images with extra noise or a different Reynolds number without updating the model parameters. |
| first_indexed | 2024-03-09T12:38:15Z |
| format | Article |
| id | doaj.art-c5d1c89ba1784aafac2d51b72d6aa37e |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-09T12:38:15Z |
| publishDate | 2022-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-c5d1c89ba1784aafac2d51b72d6aa37e2023-11-30T22:21:48ZengMDPI AGEnergies1996-10732022-08-011515574710.3390/en15155747Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh ImagingMinnan Cai0Hua Jin1Beichen Lin2Wenjiang Xu3Yancheng You4School of Aerospace Engineering, Xiamen University, Xiamen 361005, ChinaSchool of Aerospace Engineering, Xiamen University, Xiamen 361005, ChinaSchool of Aerospace Engineering, Xiamen University, Xiamen 361005, ChinaSchool of Aerospace Engineering, Xiamen University, Xiamen 361005, ChinaSchool of Aerospace Engineering, Xiamen University, Xiamen 361005, ChinaThe conventional denoising method in Rayleigh imaging in a general sense requires an additional hardware investment and the use of the underlying physics. This work demonstrates an alternative image denoising reconstruction model based on unsupervised learning that aims to remove Mie scattering and shot noise interference from two-dimensional (2D) Rayleigh images. The model has two generators and two discriminators whose parameters can be trained with either feature-paired or feature-unpaired data independently. The proposed network was extensively evaluated with a qualitative examination and quantitative metrics, such as PSNR, ER, and SSIM. The results demonstrate that the feature-paired training network exhibits a better performance compared with several other networks reported in the literature. Moreover, when the flame features are not paired, the feature-unpaired training network still yields a good agreement with ground truth data. Specific indicators of the quantitative evaluation show a promising denoising ability with a peak signal-to-noise ratio of ~37 dB, an overall reconstruction error of ~1%, and a structure similarity index of ~0.985. Additionally, the pre-trained unsupervised model based on unpaired training can be generalized to denoise Rayleigh images with extra noise or a different Reynolds number without updating the model parameters.https://www.mdpi.com/1996-1073/15/15/5747unsupervised learningnoise reductionRayleigh imaging |
| spellingShingle | Minnan Cai Hua Jin Beichen Lin Wenjiang Xu Yancheng You Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging Energies unsupervised learning noise reduction Rayleigh imaging |
| title | Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging |
| title_full | Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging |
| title_fullStr | Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging |
| title_full_unstemmed | Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging |
| title_short | Numerical Demonstration of Unsupervised-Learning-Based Noise Reduction in Two-Dimensional Rayleigh Imaging |
| title_sort | numerical demonstration of unsupervised learning based noise reduction in two dimensional rayleigh imaging |
| topic | unsupervised learning noise reduction Rayleigh imaging |
| url | https://www.mdpi.com/1996-1073/15/15/5747 |
| work_keys_str_mv | AT minnancai numericaldemonstrationofunsupervisedlearningbasednoisereductionintwodimensionalrayleighimaging AT huajin numericaldemonstrationofunsupervisedlearningbasednoisereductionintwodimensionalrayleighimaging AT beichenlin numericaldemonstrationofunsupervisedlearningbasednoisereductionintwodimensionalrayleighimaging AT wenjiangxu numericaldemonstrationofunsupervisedlearningbasednoisereductionintwodimensionalrayleighimaging AT yanchengyou numericaldemonstrationofunsupervisedlearningbasednoisereductionintwodimensionalrayleighimaging |