This paper proposes a detection method based on an improved Mask Region-based Convolutional Neural Network(Mask R-CNN)model for crack recognition in shallow-buried compressed air energy storage(CAES)cavern linings,ena...This paper proposes a detection method based on an improved Mask Region-based Convolutional Neural Network(Mask R-CNN)model for crack recognition in shallow-buried compressed air energy storage(CAES)cavern linings,enabling a comprehensive safety assessment of gas storage caverns.Flexible concrete samples are prepared to simulate the crack characteristics of the sealing lining,providing data support for the recognition module.The Convolutional Block Attention Module is introduced into the ResNet-50 backbone to adaptively adjust feature map weights and enhance feature extraction.Additionally,the mask segmentation loss function is optimized by combining Binary Cross-Entropy loss and Dice loss to improve crack region recognition.Experimental results show that the improved Mask R-CNN model achieves a mean average precision of 89.3%,a 17.2%improvement over the original model,and an intersection over union of 88.41%.Compared to RCNN,Faster R-CNN,YOLOv5,and SSD,the improved model shows superior performance with higher average precision(AP)50:95,AP50,and AP75 values in crack recognition tasks.The proposed method effectively identifies cracks in the flexible concrete sealing lining of shallow-buried CAES caverns,contributing significantly to the prevention of gas storage leaks and providing a valuable approach for the comprehensive safety assessment of CAES gas storage caverns.展开更多
Detecting pavement cracks is critical for road safety and infrastructure management.Traditional methods,relying on manual inspection and basic image processing,are time-consuming and prone to errors.Recent deep-learni...Detecting pavement cracks is critical for road safety and infrastructure management.Traditional methods,relying on manual inspection and basic image processing,are time-consuming and prone to errors.Recent deep-learning(DL)methods automate crack detection,but many still struggle with variable crack patterns and environmental conditions.This study aims to address these limitations by introducing the Masker Transformer,a novel hybrid deep learning model that integrates the precise localization capabilities of Mask Region-based Convolutional Neural Network(Mask R-CNN)with the global contextual awareness of Vision Transformer(ViT).The research focuses on leveraging the strengths of both architectures to enhance segmentation accuracy and adaptability across different pavement conditions.We evaluated the performance of theMaskerTransformer against other state-of-theartmodels such asU-Net,TransformerU-Net(TransUNet),U-NetTransformer(UNETr),SwinU-NetTransformer(Swin-UNETr),You Only Look Once version 8(YoloV8),and Mask R-CNN using two benchmark datasets:Crack500 and DeepCrack.The findings reveal that the MaskerTransformer significantly outperforms the existing models,achieving the highest Dice SimilarityCoefficient(DSC),precision,recall,and F1-Score across both datasets.Specifically,the model attained a DSC of 80.04%on Crack500 and 91.37%on DeepCrack,demonstrating superior segmentation accuracy and reliability.The high precision and recall rates further substantiate its effectiveness in real-world applications,suggesting that the Masker Transformer can serve as a robust tool for automated pavement crack detection,potentially replacing more traditional methods.展开更多
M50轴承钢中主要的碳化物类型为MC、M_(2)C和M_(23)C_(6)。扫描电子显微镜(Scanning Electron Microscopy,SEM)下,3种碳化物的形状、尺寸和在材料中的分布存在明显的区别。有些碳化物的尺寸较大且分布不均匀。轴承受载过程中,这些碳化...M50轴承钢中主要的碳化物类型为MC、M_(2)C和M_(23)C_(6)。扫描电子显微镜(Scanning Electron Microscopy,SEM)下,3种碳化物的形状、尺寸和在材料中的分布存在明显的区别。有些碳化物的尺寸较大且分布不均匀。轴承受载过程中,这些碳化物会成为应力集中的区域,对轴承疲劳性能产生负面影响。为了高效地获得材料中的碳化物信息,提出一种改进的掩膜基于区域的卷积神经网络(Mask Region-based Convolutional Neural Network,Mask R-CNN)模型,可批量鉴别SEM图像中3种碳化物的种类,确定其尺寸大小及分布。网络模型输出的图像和数值结果显示,M50轴承钢中M_(2)C型碳化物尺寸大且分布不均匀,但总体尺寸最大的MC型碳化物和尺寸最小的M_(23)C_(6)型碳化物分布相对均匀。展开更多
基金the financial supports of the National Natural Science Foundation of China Youth Science Foundation Project(52204152,52204111,52204153)the Postdoctoral Innovation Talent Support Program(BX2020275)the Postdoctoral Science Foundation(2020M683521).
文摘This paper proposes a detection method based on an improved Mask Region-based Convolutional Neural Network(Mask R-CNN)model for crack recognition in shallow-buried compressed air energy storage(CAES)cavern linings,enabling a comprehensive safety assessment of gas storage caverns.Flexible concrete samples are prepared to simulate the crack characteristics of the sealing lining,providing data support for the recognition module.The Convolutional Block Attention Module is introduced into the ResNet-50 backbone to adaptively adjust feature map weights and enhance feature extraction.Additionally,the mask segmentation loss function is optimized by combining Binary Cross-Entropy loss and Dice loss to improve crack region recognition.Experimental results show that the improved Mask R-CNN model achieves a mean average precision of 89.3%,a 17.2%improvement over the original model,and an intersection over union of 88.41%.Compared to RCNN,Faster R-CNN,YOLOv5,and SSD,the improved model shows superior performance with higher average precision(AP)50:95,AP50,and AP75 values in crack recognition tasks.The proposed method effectively identifies cracks in the flexible concrete sealing lining of shallow-buried CAES caverns,contributing significantly to the prevention of gas storage leaks and providing a valuable approach for the comprehensive safety assessment of CAES gas storage caverns.
文摘Detecting pavement cracks is critical for road safety and infrastructure management.Traditional methods,relying on manual inspection and basic image processing,are time-consuming and prone to errors.Recent deep-learning(DL)methods automate crack detection,but many still struggle with variable crack patterns and environmental conditions.This study aims to address these limitations by introducing the Masker Transformer,a novel hybrid deep learning model that integrates the precise localization capabilities of Mask Region-based Convolutional Neural Network(Mask R-CNN)with the global contextual awareness of Vision Transformer(ViT).The research focuses on leveraging the strengths of both architectures to enhance segmentation accuracy and adaptability across different pavement conditions.We evaluated the performance of theMaskerTransformer against other state-of-theartmodels such asU-Net,TransformerU-Net(TransUNet),U-NetTransformer(UNETr),SwinU-NetTransformer(Swin-UNETr),You Only Look Once version 8(YoloV8),and Mask R-CNN using two benchmark datasets:Crack500 and DeepCrack.The findings reveal that the MaskerTransformer significantly outperforms the existing models,achieving the highest Dice SimilarityCoefficient(DSC),precision,recall,and F1-Score across both datasets.Specifically,the model attained a DSC of 80.04%on Crack500 and 91.37%on DeepCrack,demonstrating superior segmentation accuracy and reliability.The high precision and recall rates further substantiate its effectiveness in real-world applications,suggesting that the Masker Transformer can serve as a robust tool for automated pavement crack detection,potentially replacing more traditional methods.
