Segmentation of computed tomography images and high-precision reconstruction of rubber composite structure based on deep learning
Source abstract
Abstract
Abstract The internal fiber structure of composites determined via X-ray micro-computed tomography (μ-CT) is a key factor affecting their properties. However, the low resolution and blurred fiber-matrix interfaces of the μ-CT images obtained for large composite structures as well as the different grayscale distributions caused by the uneven thicknesses of their complex shapes considerably reduce the accuracy of the reconstructed model. Meanwhile, the relatively long scanning time of μ-CT limits the usability of this technique for in-situ loading tests. Therefore, in this study, an image segmentation method based on U-net convolution neural network is applied to the segmentation of μ-CT image of fabric rubber composite to realize the high-precision reconstruction of its internal fabric structure. This approach effectively minimizes the image noise caused by fast scanning and reduces the scanning time by 80%. The accuracy of the reconstructed model and the high efficiency of the developed method are verified through finite element analysis. Hence, the described μ-CT image processing method enables high-precision reconstruction of fiber-reinforced composites and provides a significant reduction in the μ-CT scanning time.
Source: openaire · Quality: candidate unverified · Source record ↗
Research context
OpenAlex topics: Medical Imaging Techniques and Applications · Advanced X-ray and CT Imaging · Advanced X-ray Imaging Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (40)
Show source-provided references
1. Carmignato · 2018
2. In situ fibre fracture measurement in carbon–epoxy laminates using high resolution computed tomography10.1016/j.compscitech.2011.06.004 ↗
3. In situ X-ray micro-computed tomography study of the damage evolution of prefabricated through-holes in SLM-Printed AlSi10Mg alloy under tension10.1016/j.jallcom.2019.153576 ↗
4. Damage mechanisms in elastomeric foam composites: multiscale X-ray computed tomography and finite element analyses10.1016/j.compscitech.2018.11.025 ↗
5. Effect of manufacturing defect on mechanical performance of plain weave carbon/epoxy composite based on 3D geometrical reconstruction10.1016/j.compstruct.2018.05.066 ↗
6. X-ray computed tomography of a titanium aerospace investment casting10.1016/j.csndt.2015.03.001 ↗
7. On the challenges of greyscale‐based quantifications using X-ray computed microtomography10.1111/jmi.12805 ↗
8. A review on CT image noise and its denoising10.1016/j.bspc.2018.01.010 ↗
9. 20 Hz X-ray tomography during an in situ tensile test10.1007/s10704-016-0077-y ↗
10. High resolution synchrotron-based radiography and tomography using hard X-rays at the BAMline (BESSY II)10.1016/j.nima.2007.11.020 ↗
11. Multi‐Scale imaging of polymer electrolyte fuel cells using X-ray micro- and nano-computed tomography, transmission electron microscopy and helium-ion microscopy10.1002/fuce.201800047 ↗
12. X-ray microtomography of materials10.1179/095066099101528261 ↗
13. Image reconstruction is a new frontier of machine learning10.1109/TMI.2018.2833635 ↗
14. A mixed-scale dense convolutional neural network for image analysis10.1073/pnas.1715832114 ↗
15. Improving tomographic reconstruction from limited data using mixed-scale dense convolutional neural networks10.3390/jimaging4110128 ↗
16. Deep convolutional neural network for inverse problems in imaging10.1109/TIP.2017.2713099 ↗
17. Cone beam computed tomography image quality improvement using a deep convolutional neural network
18. Ronneberger · 2015 · 234
19. H-DenseUNet: Hybrid densely connected UNet for liver and tumor segmentation from CT volumes10.1109/TMI.2018.2845918 ↗
20. Unet++: a nested u-net architecture for medical image segmentation
21. Semantic segmentation of the micro-structure of strain-hardening cement-based composites (SHCC) by applying deep learning on micro-computed tomography scans10.1016/j.cemconcomp.2020.103551 ↗
22. Automated segmentation of computed tomography images of fiber-reinforced composites by deep learning10.1007/s10853-020-05148-7 ↗
23. Microstructural analysis of short glass fiber reinforced thermoplastics based on x-ray micro-computed tomography10.1016/j.compscitech.2019.107752 ↗
24. Edge flow: a framework of boundary detection and image segmentation10.1109/CVPR.1997.609409 ↗
25. A segmentation system based on thresholding10.1016/S0146-664X(81)80015-9 ↗
26. C · 1977
27. The watershed transform: definitions, algorithms and parallelization strategies
28. Mitiche · 2010
29. Image denoising method based on curvelet transform
30. Deep learning for brain MRI segmentation: state of the art and future directions10.1007/s10278-017-9983-4 ↗
31. Lung CT image segmentation using deep neural networks10.1016/j.procs.2018.01.104 ↗
32. Advanced deep learning‐based 3D microstructural characterization of multiphase metal matrix composites10.1002/adem.201901197 ↗
33. Segmenting delaminations in carbon fiber reinforced polymer composite CT using convolutional neural networks10.1063/1.4940585 ↗
34. Variational and deep learning segmentation of very-low-contrast X-ray computed tomography images of carbon/epoxy woven composites10.3390/ma13040936 ↗
35. Konopczyński · 2019
36. Constitutive behaviors and mechanical characterizations of fabric reinforced rubber composites10.1016/j.compstruct.2016.05.021 ↗
37. Macro- and mesoscopic mechanical properties of complex fabric rubber composite under different temperatures10.1016/j.compstruct.2019.111510 ↗
38. Effect of stress relaxation on sealing performance of the fabric rubber seal10.1016/j.compscitech.2017.08.025 ↗
39. Anisotropic hyper-viscoelastic behaviors of fabric reinforced rubber composites10.1016/j.compstruct.2017.12.026 ↗
40. Simulia Corp · 2017