Super-resolution single-molecule reconstruction of composite infrared images based on deep machine learning
Abstract
In order to improve the resolution of infrared images of single-molecule reconstruction of composite materials, this paper proposes an image super-resolution reconstruction method based on deep machine learning SRGAN, By replacing the residual blocks of the generated network in SRGAN with residual dense block, it can more effectively acquire and utilize the image features from various network layers, especially those containing high-frequency information, thereby ensuring that more details and textures are preserved during the magnification of infrared images. The SE attention mechanism is incorporated into the generative network by assigning a weight to each channel, which strengthens the focus on important features while reducing reliance on irrelevant information. Super-resolution reconstruction experiments conducted on CFRP composite material infrared images demonstrate that the improved algorithm achieves a 0.6 increase in Peak Signal-to-Noise Ratio (PSNR) and a 0.3% increase in Structural Similarity Index (SSIM) compared to SRGAN, providing valuable references for the super-resolution reconstruction of infrared images of composite materials.
References
1. Kupski J, de Freitas ST. Design of adhesively bonded lap joints with laminated CFRP adherends: Review, challenges and new opportunities for aerospace structures[J]. Composite Structures, 2021, 268: 113923.
2. Stoiber N, Hammerl M, Kromoser B. Cradle-to-gate life cycle assessment of CFRP reinforcement for concrete structures: Calculation basis and exemplary application[J]. Journal of Cleaner Production, 2021, 280: 124300.
3. Mahdavi M, Yousefi E, Baniassadi M, et al. Effective thermal and mechanical properties of short carbon fiber/natural rubber composites as a function of mechanical loading[J]. Applied Thermal Engineering, 2017, 117: 8–16.
4. Li D. Carbon fiber reinforced composites make the aircraft structure lighter and stronger [J]. Large Aircraft, 2020 (02): 76–79.
5. Zhao Y. Application of advanced fiber-reinforced resin-based composites in the aerospace industry [J]. Dual-use Technologies and Products, 2010 (1): 4–6.
6. Jiang H, Chen L. Application of infrared thermal wave imaging in nondestructive detection of composites [J]. Nondestructive testing, 2018,40 (11): 37–41.
7. Dou LX, Wang Q, Chen G. Study on the low-speed impact characteristics and damage analysis of carbon fiber composites [J]. Mechanical and electrical Engineering, 2016,33 (7): 815–821.
8. Lin L, Fu X, Huang X, et al. Infrared thermal imaging study for the detection of defects in composite materials [J]. Electromechanical Engineering, 2019,36 (06): 628–632.
9. Wei X, Xiong J, Wang J, Xu W. Progress in the cellular structure of fiber reinforced composites [J] Chinese Science: Technical Science, 2020,50 (08): 1123–1124.
10. Wei J, Liu J, He L, et al. Research and development status of nondestructive testing technology for infrared thermal imaging [J]. Journal of Harbin University of Science and Technology, 2020,25 (02): 64–72.
11. An S. Research on infrared thermal wave detection of CFRP laminate [D]. Heilongjiang University of Science and Technology, 2023.
12. Kong S, Xie Y, Wang S, et al. Review of the development of infrared thermal image enhancement algorithm [J]. Journal of Chongqing University of Science and Technology (Natural Science Edition), 2021,23 (04): 77–83.
13. Harris JL. Diffraction and Resolving Power [J]. Journal of the Optical Society of America, 1964, 54(7): 931–933.
14. Dong C, Loy C, He K, et al. Image super-resolution using deep convolutional networks[J]. IEEE transactions on pattern analysis and machine intelligence(S0162–8828), 2015, 38(2): 295–307.
15. Ledig C, Theis L, Huszar F, et al. Photo-realistic single image super-resolution using a generative adversarial network[C]// Proceedings of the IEEE conference on computer vision and pattern recognition.2017:4681–4690.
16. Wu T, Yang W, Tao X, et al. Super-resolution reconstruction of infrared images of transmission line insulators based on a lightweight, dense residual network [J / OL]. High-voltage electrical appliances, 1–12 [2024-06-27].
17. Liu Z, Tao Y, Liu H, et al. Super-resolution reconstruction of infrared patrol images incorporating residual dense and generative adversarial networks [J]. Journal of Kunming University of Science and Technology (Natural Science edition), 2023,48 (05): 120–129.
18. Hu D, Min T. Infrared image super-resolution method for a modified lightweight GAN [J]. Small Microcomputer system, 2022,43 (08): 1711–1717.
19. Zhang Y, Tian Y, Kong Y, et al. Residual Dense Network for Image Super-Resolution[C]//IEEE Conference on Computer Vision & Pattern Recognition.2018.
Copyright (c) 2025 Author(s)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright on all articles published in this journal is retained by the author(s), while the author(s) grant the publisher as the original publisher to publish the article.
Articles published in this journal are licensed under a Creative Commons Attribution 4.0 International, which means they can be shared, adapted and distributed provided that the original published version is cited.
Submit a Paper
