Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology
Fiber extraction and segmentation are critical to identifying the distribution of short fibers in fiber reinforced concrete. This work presents a new segmentation model based on the DeepLab V3 + network for automated probabilistic segmentation of fibers embedded in glass fiber reinforced concrete (G...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2021-11-01
|
Series: | Materials & Design |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S0264127521005797 |
_version_ | 1798035190815129600 |
---|---|
author | Li Hong Peng Zhang Dongxu Liu Peng Gao Binggen Zhan Qijun Yu Lizhi Sun |
author_facet | Li Hong Peng Zhang Dongxu Liu Peng Gao Binggen Zhan Qijun Yu Lizhi Sun |
author_sort | Li Hong |
collection | DOAJ |
description | Fiber extraction and segmentation are critical to identifying the distribution of short fibers in fiber reinforced concrete. This work presents a new segmentation model based on the DeepLab V3 + network for automated probabilistic segmentation of fibers embedded in glass fiber reinforced concrete (GFRC) images obtained from micro-computed tomography (Micro-CT). A total of 2700 sliced images (715 × 715 pixels) are prepared by Micro-CT scanning and data augmentation, of which 2400 are for training with the remaining 300 for validation. It is shown that the proposed model can accurately and effectively segment the fibers in GFRC. The efficiency of the proposed segmentation model is accessed by three metrics of accuracy, intersection over union, and F1-score index, and they are reaching up to 99.3%, 80.4% and 67.2%, respectively. Moreover, the model can achieve a better segmentation than the traditional deep learning models. With good qualitative prediction, the proposed approach shows promise for predicting fiber segmentation based on unlabeled data obtained in the field. Finally, three-dimensional distribution of short fibers in GFRC samples with size of 2.99 × 2.99 × 2.99 mm3 is developed for reconstruction analysis. |
first_indexed | 2024-04-11T20:54:48Z |
format | Article |
id | doaj.art-7c1265493eaf420080665a4e2db859fb |
institution | Directory Open Access Journal |
issn | 0264-1275 |
language | English |
last_indexed | 2024-04-11T20:54:48Z |
publishDate | 2021-11-01 |
publisher | Elsevier |
record_format | Article |
series | Materials & Design |
spelling | doaj.art-7c1265493eaf420080665a4e2db859fb2022-12-22T04:03:43ZengElsevierMaterials & Design0264-12752021-11-01210110024Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technologyLi Hong0Peng Zhang1Dongxu Liu2Peng Gao3Binggen Zhan4Qijun Yu5Lizhi Sun6Department of Structural Engineering, Hefei University of Technology, Hefei 230009, China; Key Laboratory of Performance Evolution and Control for Engineering Structures, Tongji University, Shanghai 200092, China; Department of Civil & Environmental Engineering, University of California, Irvine, CA 92697, USA; Hefei Cement Research & Design Institute Corporation Ltd, Hefei 230051, ChinaDepartment of Structural Engineering, Hefei University of Technology, Hefei 230009, ChinaDepartment of Civil & Environmental Engineering, University of California, Irvine, CA 92697, USADepartment of Structural Engineering, Hefei University of Technology, Hefei 230009, ChinaDepartment of Structural Engineering, Hefei University of Technology, Hefei 230009, ChinaDepartment of Structural Engineering, Hefei University of Technology, Hefei 230009, China; Corresponding authors.Department of Civil & Environmental Engineering, University of California, Irvine, CA 92697, USA; Corresponding authors.Fiber extraction and segmentation are critical to identifying the distribution of short fibers in fiber reinforced concrete. This work presents a new segmentation model based on the DeepLab V3 + network for automated probabilistic segmentation of fibers embedded in glass fiber reinforced concrete (GFRC) images obtained from micro-computed tomography (Micro-CT). A total of 2700 sliced images (715 × 715 pixels) are prepared by Micro-CT scanning and data augmentation, of which 2400 are for training with the remaining 300 for validation. It is shown that the proposed model can accurately and effectively segment the fibers in GFRC. The efficiency of the proposed segmentation model is accessed by three metrics of accuracy, intersection over union, and F1-score index, and they are reaching up to 99.3%, 80.4% and 67.2%, respectively. Moreover, the model can achieve a better segmentation than the traditional deep learning models. With good qualitative prediction, the proposed approach shows promise for predicting fiber segmentation based on unlabeled data obtained in the field. Finally, three-dimensional distribution of short fibers in GFRC samples with size of 2.99 × 2.99 × 2.99 mm3 is developed for reconstruction analysis.http://www.sciencedirect.com/science/article/pii/S0264127521005797Glass fiber reinforced concreteDeep learningShort fiberMicro-CTFiber segmentation |
spellingShingle | Li Hong Peng Zhang Dongxu Liu Peng Gao Binggen Zhan Qijun Yu Lizhi Sun Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology Materials & Design Glass fiber reinforced concrete Deep learning Short fiber Micro-CT Fiber segmentation |
title | Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology |
title_full | Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology |
title_fullStr | Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology |
title_full_unstemmed | Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology |
title_short | Effective segmentation of short fibers in glass fiber reinforced concrete’s X-ray images using deep learning technology |
title_sort | effective segmentation of short fibers in glass fiber reinforced concrete s x ray images using deep learning technology |
topic | Glass fiber reinforced concrete Deep learning Short fiber Micro-CT Fiber segmentation |
url | http://www.sciencedirect.com/science/article/pii/S0264127521005797 |
work_keys_str_mv | AT lihong effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT pengzhang effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT dongxuliu effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT penggao effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT binggenzhan effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT qijunyu effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology AT lizhisun effectivesegmentationofshortfibersinglassfiberreinforcedconcretesxrayimagesusingdeeplearningtechnology |