Sheng Zhang , Kaiyu Wang , Huajun Zhang , Tong Wang , Xiguang Gao , Yingdong Song , Fang Wang
{"title":"改进 YOLOv8,用于在 2.5D 复合材料的 X 射线计算机断层扫描图像中分割纤维束,以建立有限元模型","authors":"Sheng Zhang , Kaiyu Wang , Huajun Zhang , Tong Wang , Xiguang Gao , Yingdong Song , Fang Wang","doi":"10.1016/j.compositesa.2024.108337","DOIUrl":null,"url":null,"abstract":"<div><p>It is necessary to segment fiber bundles in the reconstruction of the mesoscopic model of ceramic matrix composites using XCT images. Existing methods have great subjectivity, poor recognition accuracy, and heavy workload. To solve this problem, an improved lightweight YOLOv8 was proposed, which is a deep learning approach. By adding Slim-neck and VanillaNet, the complexity of the model was greatly reduced. Additionally, by replacing the loss function of the model with the Wise-IoU loss function, the ability of feature extraction of the model was improved. The effectiveness of the improved YOLOv8 in fiber bundle identification was demonstrated. Finally, a mesoscopic model was reconstructed by XCT images where fiber bundles were segmented by using the improved YOLOv8. The linear elastic modulus of the material was predicted and the error was found to be small, indicating that the improved YOLOv8 can effectively segment fiber bundles and thus reconstruct a high-precision mesoscopic model.</p></div>","PeriodicalId":282,"journal":{"name":"Composites Part A: Applied Science and Manufacturing","volume":null,"pages":null},"PeriodicalIF":8.1000,"publicationDate":"2024-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An improved YOLOv8 for fiber bundle segmentation in X-ray computed tomography images of 2.5D composites to build the finite element model\",\"authors\":\"Sheng Zhang , Kaiyu Wang , Huajun Zhang , Tong Wang , Xiguang Gao , Yingdong Song , Fang Wang\",\"doi\":\"10.1016/j.compositesa.2024.108337\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>It is necessary to segment fiber bundles in the reconstruction of the mesoscopic model of ceramic matrix composites using XCT images. Existing methods have great subjectivity, poor recognition accuracy, and heavy workload. To solve this problem, an improved lightweight YOLOv8 was proposed, which is a deep learning approach. By adding Slim-neck and VanillaNet, the complexity of the model was greatly reduced. Additionally, by replacing the loss function of the model with the Wise-IoU loss function, the ability of feature extraction of the model was improved. The effectiveness of the improved YOLOv8 in fiber bundle identification was demonstrated. Finally, a mesoscopic model was reconstructed by XCT images where fiber bundles were segmented by using the improved YOLOv8. The linear elastic modulus of the material was predicted and the error was found to be small, indicating that the improved YOLOv8 can effectively segment fiber bundles and thus reconstruct a high-precision mesoscopic model.</p></div>\",\"PeriodicalId\":282,\"journal\":{\"name\":\"Composites Part A: Applied Science and Manufacturing\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":8.1000,\"publicationDate\":\"2024-06-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Composites Part A: Applied Science and Manufacturing\",\"FirstCategoryId\":\"1\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1359835X24003348\",\"RegionNum\":2,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, MANUFACTURING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Composites Part A: Applied Science and Manufacturing","FirstCategoryId":"1","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1359835X24003348","RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MANUFACTURING","Score":null,"Total":0}
An improved YOLOv8 for fiber bundle segmentation in X-ray computed tomography images of 2.5D composites to build the finite element model
It is necessary to segment fiber bundles in the reconstruction of the mesoscopic model of ceramic matrix composites using XCT images. Existing methods have great subjectivity, poor recognition accuracy, and heavy workload. To solve this problem, an improved lightweight YOLOv8 was proposed, which is a deep learning approach. By adding Slim-neck and VanillaNet, the complexity of the model was greatly reduced. Additionally, by replacing the loss function of the model with the Wise-IoU loss function, the ability of feature extraction of the model was improved. The effectiveness of the improved YOLOv8 in fiber bundle identification was demonstrated. Finally, a mesoscopic model was reconstructed by XCT images where fiber bundles were segmented by using the improved YOLOv8. The linear elastic modulus of the material was predicted and the error was found to be small, indicating that the improved YOLOv8 can effectively segment fiber bundles and thus reconstruct a high-precision mesoscopic model.
期刊介绍:
Composites Part A: Applied Science and Manufacturing is a comprehensive journal that publishes original research papers, review articles, case studies, short communications, and letters covering various aspects of composite materials science and technology. This includes fibrous and particulate reinforcements in polymeric, metallic, and ceramic matrices, as well as 'natural' composites like wood and biological materials. The journal addresses topics such as properties, design, and manufacture of reinforcing fibers and particles, novel architectures and concepts, multifunctional composites, advancements in fabrication and processing, manufacturing science, process modeling, experimental mechanics, microstructural characterization, interfaces, prediction and measurement of mechanical, physical, and chemical behavior, and performance in service. Additionally, articles on economic and commercial aspects, design, and case studies are welcomed. All submissions undergo rigorous peer review to ensure they contribute significantly and innovatively, maintaining high standards for content and presentation. The editorial team aims to expedite the review process for prompt publication.