Pub Date : 2025-02-13DOI: 10.1109/TIM.2025.3533640
Yifan Wei;Zelong Ma;Handing Xu;Yanjie Xu;Deli Chen;Yanjin Dong;Zhenguo Nie
The landslide incidence around petroleum pipelines usually leads to severe pipeline damage, causing serious environmental pollution, economic losses, and even casualties. The commonly used methods for monitoring petroleum pipelines include remote sensing and ground monitoring, which can be further applied in predicting landslide hazards. However, current landslide prediction strategies are generally limited because the provided predictive indicators are restricted, and the prediction needs to be more timely. These limitations have caused significant difficulties in the practical application of landslide prediction. Based on the actual operating conditions and landslide incidence records of a section of the pipeline, we propose a machine learning-based landslide prewarning system for buried pipelines to achieve real-time landslide monitoring and rapid warning. Landslide conditions and pipeline operation records are collected using a ground monitoring system around the target pipeline section. They are integrated to generate machine learning datasets extended using the finite element method (FEM). After comparing multiple machine learning algorithms, the XGBoost model is ultimately adopted for the prediction system. The system is calibrated and verified by comparing the prediction results with landslide data. In the verification test, the landslide warning message is acquired about 10 h before the landslide occurrence, and the error in predicting the position of the landslide is about 20 m in the case of the 466 m target pipeline.
{"title":"Real-Time Prewarning System for Petroleum Pipeline Landslide Prediction Based on Imbalanced Machine Learning Methods With Finite Element Analysis","authors":"Yifan Wei;Zelong Ma;Handing Xu;Yanjie Xu;Deli Chen;Yanjin Dong;Zhenguo Nie","doi":"10.1109/TIM.2025.3533640","DOIUrl":"https://doi.org/10.1109/TIM.2025.3533640","url":null,"abstract":"The landslide incidence around petroleum pipelines usually leads to severe pipeline damage, causing serious environmental pollution, economic losses, and even casualties. The commonly used methods for monitoring petroleum pipelines include remote sensing and ground monitoring, which can be further applied in predicting landslide hazards. However, current landslide prediction strategies are generally limited because the provided predictive indicators are restricted, and the prediction needs to be more timely. These limitations have caused significant difficulties in the practical application of landslide prediction. Based on the actual operating conditions and landslide incidence records of a section of the pipeline, we propose a machine learning-based landslide prewarning system for buried pipelines to achieve real-time landslide monitoring and rapid warning. Landslide conditions and pipeline operation records are collected using a ground monitoring system around the target pipeline section. They are integrated to generate machine learning datasets extended using the finite element method (FEM). After comparing multiple machine learning algorithms, the XGBoost model is ultimately adopted for the prediction system. The system is calibrated and verified by comparing the prediction results with landslide data. In the verification test, the landslide warning message is acquired about 10 h before the landslide occurrence, and the error in predicting the position of the landslide is about 20 m in the case of the 466 m target pipeline.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-8"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Evaluating measurement uncertainty is crucial for ensuring the reliability of piezoelectric drive systems. However, existing international standards are insufficient for dynamic measurement uncertainty evaluation, primarily due to the complexity of dynamic systems and the challenges of establishing uncertainty propagation models with limited samples. To address this issue, we propose a temporal evidential regression network (T-ENet) for developing dynamic models and evaluating uncertainty in piezoelectric impulse drive systems with small-sample nonlinear characteristics. We combine an evidential regression model with gated recurrent units (GRUs) to create a robust modeling framework. This framework integrates gradient-updated meta-learning algorithms, allowing it to perform effectively with minimal training data and gradient updates, accurately capturing the temporal features of dynamic systems and estimating and predicting the distribution parameters of system dynamic uncertainty. Experimental results validate the effectiveness of our method, and comparisons with traditional long short-term memory (LSTM) and GRU networks demonstrate its superiority in dynamic prediction. The correlation between prediction uncertainty and actual error confirms the effectiveness of our method in estimating uncertainty in dynamic measurements and provides a key reference for analyzing the reliability of actual measurement results.
{"title":"Dynamic Modeling and Measurement Uncertainty Evaluation of the Nonlinear Piezoelectric Impulse Drive System With Small Samples","authors":"Yinye Ding;Wenhao Chen;Rencheng Song;Hongli Li;Chengliang Pan;Haojie Xia","doi":"10.1109/TIM.2025.3541670","DOIUrl":"https://doi.org/10.1109/TIM.2025.3541670","url":null,"abstract":"Evaluating measurement uncertainty is crucial for ensuring the reliability of piezoelectric drive systems. However, existing international standards are insufficient for dynamic measurement uncertainty evaluation, primarily due to the complexity of dynamic systems and the challenges of establishing uncertainty propagation models with limited samples. To address this issue, we propose a temporal evidential regression network (T-ENet) for developing dynamic models and evaluating uncertainty in piezoelectric impulse drive systems with small-sample nonlinear characteristics. We combine an evidential regression model with gated recurrent units (GRUs) to create a robust modeling framework. This framework integrates gradient-updated meta-learning algorithms, allowing it to perform effectively with minimal training data and gradient updates, accurately capturing the temporal features of dynamic systems and estimating and predicting the distribution parameters of system dynamic uncertainty. Experimental results validate the effectiveness of our method, and comparisons with traditional long short-term memory (LSTM) and GRU networks demonstrate its superiority in dynamic prediction. The correlation between prediction uncertainty and actual error confirms the effectiveness of our method in estimating uncertainty in dynamic measurements and provides a key reference for analyzing the reliability of actual measurement results.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-11"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-13DOI: 10.1109/TIM.2025.3541666
Jin Bao;Zhengye Shen;Guisong Chen;Xuecheng Zhao;Zengwang Yang
With the rapid advancement of industrialization and urbanization, the adverse effects of air pollution on human health and environmental protection have become increasingly significant. This study developed an air quality monitoring device equipped with various air detection sensors and integrated with a Wi-Fi sensor for data collection and cloud upload. A multilayer long short-term memory (LSTM) model was used to analyze the data, and strategies for deployment on edge computing devices were explored. The study also leveraged the high performance and low power consumption of embedded chips to process air quality data locally in real time. Experimental results showed that the system achieved 91.6% accuracy. In terms of precision and accuracy, our model improved by 8.3% and 10.6%, respectively, compared to traditional multilayer perceptron (MLP) and by 9.7% and 11.3%, respectively, compared to recurrent neural network (RNN), significantly enhancing the efficiency and reliability of air quality classification. Moreover, this research not only provides new perspectives for environmental monitoring and data processing but also elucidates the application of edge computing in intelligent environmental monitoring, which is crucial for promoting low-carbon development.
{"title":"Intelligent Air Quality Detection Device Based on Edge Computing","authors":"Jin Bao;Zhengye Shen;Guisong Chen;Xuecheng Zhao;Zengwang Yang","doi":"10.1109/TIM.2025.3541666","DOIUrl":"https://doi.org/10.1109/TIM.2025.3541666","url":null,"abstract":"With the rapid advancement of industrialization and urbanization, the adverse effects of air pollution on human health and environmental protection have become increasingly significant. This study developed an air quality monitoring device equipped with various air detection sensors and integrated with a Wi-Fi sensor for data collection and cloud upload. A multilayer long short-term memory (LSTM) model was used to analyze the data, and strategies for deployment on edge computing devices were explored. The study also leveraged the high performance and low power consumption of embedded chips to process air quality data locally in real time. Experimental results showed that the system achieved 91.6% accuracy. In terms of precision and accuracy, our model improved by 8.3% and 10.6%, respectively, compared to traditional multilayer perceptron (MLP) and by 9.7% and 11.3%, respectively, compared to recurrent neural network (RNN), significantly enhancing the efficiency and reliability of air quality classification. Moreover, this research not only provides new perspectives for environmental monitoring and data processing but also elucidates the application of edge computing in intelligent environmental monitoring, which is crucial for promoting low-carbon development.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-10"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shearography, a commonly utilized method for nondestructive testing (NDT), offers the advantage of full-field, noncontact, and real-time inspection across extensive viewing areas. Nonetheless, the effectiveness of defect detection using this method is frequently compromised by background deformation, posing challenges in accurately determining defect sizes and locations. This study investigates the deformation observed in specimens during shearography defect detection and introduces a background fringe suppression strategy based on deep learning. The aforementioned approach, which incorporates a robust dataset generation methodology and an advanced network architecture, excels at processing noisy wrapped phase maps, thereby efficiently discerning and mitigating background fringes. The effectiveness of the proposed method is corroborated through simulated and actual defect detection experiments, underscoring its potential in enhancing the precision of shearography defect detection.
{"title":"Phase Fringe Suppression of Background Deformation for Nondestructive Testing Based on Shearography","authors":"Yonghong Wang;Zihua Zheng;Xiangwei Liu;Peizheng Yan;Junrui Li;Zhenmin Zhu","doi":"10.1109/TIM.2025.3541693","DOIUrl":"https://doi.org/10.1109/TIM.2025.3541693","url":null,"abstract":"Shearography, a commonly utilized method for nondestructive testing (NDT), offers the advantage of full-field, noncontact, and real-time inspection across extensive viewing areas. Nonetheless, the effectiveness of defect detection using this method is frequently compromised by background deformation, posing challenges in accurately determining defect sizes and locations. This study investigates the deformation observed in specimens during shearography defect detection and introduces a background fringe suppression strategy based on deep learning. The aforementioned approach, which incorporates a robust dataset generation methodology and an advanced network architecture, excels at processing noisy wrapped phase maps, thereby efficiently discerning and mitigating background fringes. The effectiveness of the proposed method is corroborated through simulated and actual defect detection experiments, underscoring its potential in enhancing the precision of shearography defect detection.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-7"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-13DOI: 10.1109/TIM.2025.3540143
Alexandra Karpilow;Asja Derviškadić;Mario Paolone
With the integration of distributed energy resources and the trend toward low-inertia power grids, the frequency and severity of grid dynamics are expected to increase. Conventional phasor-based signal-processing methods are proving to be insufficient in the analysis of nonstationary ac voltage and current waveforms, while the computational complexity of many dynamic signal analysis techniques hinders their deployment in operational embedded systems. This article presents the functional basis analysis (FBA), a signal-processing tool capable of capturing the broadband nature of common single-component signal dynamics in power grids while maintaining a streamlined design for real-time monitoring applications. Relying on the Hilbert transform (HT) and optimization techniques, the FBA can be user-engineered to identify and characterize combinations of several of the most common signal dynamics in power grids, including amplitude/phase modulations (AMs/PMs), frequency ramps (FRs), and steps. This article describes the theoretical basis and design of the FBA as well as the deployment of the algorithm in embedded hardware systems, with adaptations made to consider latency requirements, finite memory capacity, and fixed-point precision arithmetic. For validation, a phasor measurement unit (PMU) calibrator is used to evaluate and compare the algorithm’s performance to state-of-the-art static and dynamic phasor methods. The test results highlight the potential of the FBA method for implementation in embedded systems to enhance grid situational awareness during critical grid events. Future work will investigate the extraction of multicomponent broadband signals with empirical mode decomposition (EMD) for harmonic analysis.
{"title":"Functional Basis Analysis for the Characterization of Power System Signal Dynamics: Formulation, Implementation, and Validation","authors":"Alexandra Karpilow;Asja Derviškadić;Mario Paolone","doi":"10.1109/TIM.2025.3540143","DOIUrl":"https://doi.org/10.1109/TIM.2025.3540143","url":null,"abstract":"With the integration of distributed energy resources and the trend toward low-inertia power grids, the frequency and severity of grid dynamics are expected to increase. Conventional phasor-based signal-processing methods are proving to be insufficient in the analysis of nonstationary ac voltage and current waveforms, while the computational complexity of many dynamic signal analysis techniques hinders their deployment in operational embedded systems. This article presents the functional basis analysis (FBA), a signal-processing tool capable of capturing the broadband nature of common single-component signal dynamics in power grids while maintaining a streamlined design for real-time monitoring applications. Relying on the Hilbert transform (HT) and optimization techniques, the FBA can be user-engineered to identify and characterize combinations of several of the most common signal dynamics in power grids, including amplitude/phase modulations (AMs/PMs), frequency ramps (FRs), and steps. This article describes the theoretical basis and design of the FBA as well as the deployment of the algorithm in embedded hardware systems, with adaptations made to consider latency requirements, finite memory capacity, and fixed-point precision arithmetic. For validation, a phasor measurement unit (PMU) calibrator is used to evaluate and compare the algorithm’s performance to state-of-the-art static and dynamic phasor methods. The test results highlight the potential of the FBA method for implementation in embedded systems to enhance grid situational awareness during critical grid events. Future work will investigate the extraction of multicomponent broadband signals with empirical mode decomposition (EMD) for harmonic analysis.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-14"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-13DOI: 10.1109/TIM.2025.3541790
Hua Li;Mingyue Zhang
Superconducting quantum interference device (SQUID) magnetogastrogram (MGG) is a medical functional imaging method with great clinical potential for noninvasive diagnosis of gastric diseases. MGG signal frequency is about 0.05 Hz, and the low-frequency environmental noise interference is serious, can be several times stronger in magnitude than the signals of interest, and may severely impede the extraction of relevant information. Wiener filter is one classic denoising solution for biomagnetic applications. In this article, a new high-pass Wiener filter and signal processing framework for MGG measurement is proposed, which can filter low-frequency noises not only specific artifacts. The filter was successfully applied to MGG signal denoising. Using this general Wiener filter frame, the filter signal-to-noise ratio (SNR) is 11.3 dB better than the classical Wiener filter and it also has 16.7 dB of SNR better than without signal noise separation step. Based on our methods, 36-point array MGG signals were detected successfully and the results were consistent with the main gastric slow wave activity.
{"title":"One New Designed Wiener Filter Method for SQUID Magnetogastrogram Detection","authors":"Hua Li;Mingyue Zhang","doi":"10.1109/TIM.2025.3541790","DOIUrl":"https://doi.org/10.1109/TIM.2025.3541790","url":null,"abstract":"Superconducting quantum interference device (SQUID) magnetogastrogram (MGG) is a medical functional imaging method with great clinical potential for noninvasive diagnosis of gastric diseases. MGG signal frequency is about 0.05 Hz, and the low-frequency environmental noise interference is serious, can be several times stronger in magnitude than the signals of interest, and may severely impede the extraction of relevant information. Wiener filter is one classic denoising solution for biomagnetic applications. In this article, a new high-pass Wiener filter and signal processing framework for MGG measurement is proposed, which can filter low-frequency noises not only specific artifacts. The filter was successfully applied to MGG signal denoising. Using this general Wiener filter frame, the filter signal-to-noise ratio (SNR) is 11.3 dB better than the classical Wiener filter and it also has 16.7 dB of SNR better than without signal noise separation step. Based on our methods, 36-point array MGG signals were detected successfully and the results were consistent with the main gastric slow wave activity.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-11"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143489258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-13DOI: 10.1109/TIM.2025.3541755
Ge Guo;Yong Yan;Yonghui Hu;Wenbiao Zhang
Spontaneous heating and self-ignition of biomass fuels pose critical challenges in the regular operation of biomass power stations. The measurement of temperature distribution in stored biomass at power stations is essential for obtaining timely temperature information and hence minimizing fire risks. This article presents a methodology through acoustic tomography (AT) for measuring the temperature and temperature distribution of stored biomass in a silo. A measurement model is developed to obtain the time of flight in stored biomass based on cross correlation and thresholding methods. Image algorithms are utilized to visualize the temperature distribution. Wood pellets, as the primary fuel source for biomass power stations, are used as a test biomass fuel. Experimental results suggest that AT is able to measure the temperature distribution of wood pellets with an accuracy of ±6% over the range of $20~^{circ }$ C$sim 65~^{circ }$ C. In addition, the relative error of the central location in the high-temperature region is within ±4%, while the relative error in measuring the area of the high-temperature region is within ±5%.
{"title":"Measurement of 2-D Temperature Distribution in a Biomass Silo Through Acoustic Tomography","authors":"Ge Guo;Yong Yan;Yonghui Hu;Wenbiao Zhang","doi":"10.1109/TIM.2025.3541755","DOIUrl":"https://doi.org/10.1109/TIM.2025.3541755","url":null,"abstract":"Spontaneous heating and self-ignition of biomass fuels pose critical challenges in the regular operation of biomass power stations. The measurement of temperature distribution in stored biomass at power stations is essential for obtaining timely temperature information and hence minimizing fire risks. This article presents a methodology through acoustic tomography (AT) for measuring the temperature and temperature distribution of stored biomass in a silo. A measurement model is developed to obtain the time of flight in stored biomass based on cross correlation and thresholding methods. Image algorithms are utilized to visualize the temperature distribution. Wood pellets, as the primary fuel source for biomass power stations, are used as a test biomass fuel. Experimental results suggest that AT is able to measure the temperature distribution of wood pellets with an accuracy of ±6% over the range of <inline-formula> <tex-math>$20~^{circ }$ </tex-math></inline-formula>C<inline-formula> <tex-math>$sim 65~^{circ }$ </tex-math></inline-formula>C. In addition, the relative error of the central location in the high-temperature region is within ±4%, while the relative error in measuring the area of the high-temperature region is within ±5%.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-12"},"PeriodicalIF":5.6,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143489127","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-11DOI: 10.1109/TIM.2025.3538058
Yuxin Jin;Yunzhou Zhang;Dexing Shan;Zhifei Wu
Existing surface defect semantic segmentation methods are limited by costly annotated data and are unable to cope with new or rare defect types. Zero-shot learning offers a new possibility for addressing this issue by reducing reliance on extensive annotated data. However, methods that solely rely on image information waste the valuable experience that humans have accumulated in the field of defect detection. In this work, we propose a human-guided segmentation network (HGNet) based on CLIP, introducing human guidance to address the data scarcity and effectively leverage expert knowledge, leading to more accurate and reliable surface defect segmentation. HGNet, guided by the human-provided text, consists of two novel modules: 1) attention-based multilevel feature fusion (AMFF) which effectively integrates multilevel features using attention mechanisms to enhance the fine-grained information capture and 2) multimodal feature adaptive balancing (MFAB) which aligns and balances multimodal features through dynamic adjustment and optimization. Moreover, we extend HGNet to HGNet+ by incorporating interactive learning to correct segmentation errors with human-provided points. Our proposed method can generalize to unseen classes without additional training samples for retraining, meeting the practical needs of industrial defect detection. Extensive experiments on Defect-$4^{i}$ (and MVTec-ZSS) demonstrate that our method outperforms the state-of-the-art zero-shot methods by 5.7%/7.81% (6.57%/8.06%) and is even comparable to the performance of existing few-shot methods.
{"title":"Human-Guided Zero-Shot Surface Defect Semantic Segmentation","authors":"Yuxin Jin;Yunzhou Zhang;Dexing Shan;Zhifei Wu","doi":"10.1109/TIM.2025.3538058","DOIUrl":"https://doi.org/10.1109/TIM.2025.3538058","url":null,"abstract":"Existing surface defect semantic segmentation methods are limited by costly annotated data and are unable to cope with new or rare defect types. Zero-shot learning offers a new possibility for addressing this issue by reducing reliance on extensive annotated data. However, methods that solely rely on image information waste the valuable experience that humans have accumulated in the field of defect detection. In this work, we propose a human-guided segmentation network (HGNet) based on CLIP, introducing human guidance to address the data scarcity and effectively leverage expert knowledge, leading to more accurate and reliable surface defect segmentation. HGNet, guided by the human-provided text, consists of two novel modules: 1) attention-based multilevel feature fusion (AMFF) which effectively integrates multilevel features using attention mechanisms to enhance the fine-grained information capture and 2) multimodal feature adaptive balancing (MFAB) which aligns and balances multimodal features through dynamic adjustment and optimization. Moreover, we extend HGNet to HGNet+ by incorporating interactive learning to correct segmentation errors with human-provided points. Our proposed method can generalize to unseen classes without additional training samples for retraining, meeting the practical needs of industrial defect detection. Extensive experiments on Defect-<inline-formula> <tex-math>$4^{i}$ </tex-math></inline-formula> (and MVTec-ZSS) demonstrate that our method outperforms the state-of-the-art zero-shot methods by 5.7%/7.81% (6.57%/8.06%) and is even comparable to the performance of existing few-shot methods.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-13"},"PeriodicalIF":5.6,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396280","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-11DOI: 10.1109/TIM.2025.3538071
Shuaiyong Li;Zhang Yang;Jianxin Zeng;Chao Zhang
The dispersive properties of Lamb waves result in lower accuracy in defect imaging. Some dispersion compensation methods are proposed to enhance imaging accuracy, which rely on known dispersive curves commonly unavailable in practice. In this article, a dispersion compensation method for Lamb waves is introduced to enhance the accuracy of defect imaging. This method utilizes nonlinear group delay estimation (NGDE) and operates when known dispersion curves are unavailable. By replacing the group velocity curve with time-frequency ridges, the problem of unknown dispersion curves is addressed. The traditional Carmona method for ridge extraction is optimized using NGDE to obtain more accurate ridges. Additionally, a method is proposed to calculate compensatory phases based on the group delay (GD) at the central frequency, resulting in nondispersive single-component signals. The ridge extraction and dispersion compensation simulation results demonstrate that the proposed method outperforms the matching pursuit (MP) and dispersion and multimode orthogonal MP (DMOMP) regarding signal-to-noise ratio (SNR) and relative error (RE). Subsequently, the method is also verified by application to defect imaging of delay-and-sum (DAS), weighted DAS (WDAS), and minimum variance distortionless response (MVDR), respectively, which can effectively improve imaging performance.
{"title":"A Novel Dispersion Compensation of Lamb Waves by Nonlinear Group Delay Estimation for Defect Imaging","authors":"Shuaiyong Li;Zhang Yang;Jianxin Zeng;Chao Zhang","doi":"10.1109/TIM.2025.3538071","DOIUrl":"https://doi.org/10.1109/TIM.2025.3538071","url":null,"abstract":"The dispersive properties of Lamb waves result in lower accuracy in defect imaging. Some dispersion compensation methods are proposed to enhance imaging accuracy, which rely on known dispersive curves commonly unavailable in practice. In this article, a dispersion compensation method for Lamb waves is introduced to enhance the accuracy of defect imaging. This method utilizes nonlinear group delay estimation (NGDE) and operates when known dispersion curves are unavailable. By replacing the group velocity curve with time-frequency ridges, the problem of unknown dispersion curves is addressed. The traditional Carmona method for ridge extraction is optimized using NGDE to obtain more accurate ridges. Additionally, a method is proposed to calculate compensatory phases based on the group delay (GD) at the central frequency, resulting in nondispersive single-component signals. The ridge extraction and dispersion compensation simulation results demonstrate that the proposed method outperforms the matching pursuit (MP) and dispersion and multimode orthogonal MP (DMOMP) regarding signal-to-noise ratio (SNR) and relative error (RE). Subsequently, the method is also verified by application to defect imaging of delay-and-sum (DAS), weighted DAS (WDAS), and minimum variance distortionless response (MVDR), respectively, which can effectively improve imaging performance.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-10"},"PeriodicalIF":5.6,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143396279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-11DOI: 10.1109/TIM.2025.3538059
Gongxu Liu;Xu-An Liu;Lu Huang;Long Li;Xinbo Gao
Electromagnetic (EM) maps hold promise for providing strategic support for future localization and navigation. However, reconstruction of EM maps under sparse nodes conditions is a critical challenge, which constrains the reliable and rapid construction of EM maps. To address this issue, we absorb the concept of Moran’s index and use location index, fingerprint index, and scaling index to quantify the topology, quality, and quantity of EM nodes, respectively. Besides, three node selection criteria were used: one based on random selection, another using the K-means clustering, and the third involving affinity propagation (AP) clustering. Under each criterion, the overall EM maps were reconstructed using four representative interpolation methods: inverse distance weighting (IDW), modified Shepard’s method (MSM), Kriging algorithm (KGA), and biharmonic spline interpolation (v4). Extensive experiments were conducted to verify the representation model. Experimental results indicate that the node selection criterion based on AP clustering outperforms the K-means criterion, which itself is superior to random selection criterion. This is attributed to the fact that under the same conditions, the AP clustering node selection criterion can better take the quality and the topological relationship of EM nodes into account. Another significant finding is that with good node selection criterion and reconstruction algorithms, only approximately 0.3%–0.5% of the nodes are required to achieve the reconstruction performance that would otherwise require up to 20% or more of the nodes. The above conclusions are of good guidance for the reliable and rapid reconstruction of EM maps via sparse or limited nodes.
{"title":"Representation Model for Electromagnetic Maps Reconstruction via Sparse Nodes","authors":"Gongxu Liu;Xu-An Liu;Lu Huang;Long Li;Xinbo Gao","doi":"10.1109/TIM.2025.3538059","DOIUrl":"https://doi.org/10.1109/TIM.2025.3538059","url":null,"abstract":"Electromagnetic (EM) maps hold promise for providing strategic support for future localization and navigation. However, reconstruction of EM maps under sparse nodes conditions is a critical challenge, which constrains the reliable and rapid construction of EM maps. To address this issue, we absorb the concept of Moran’s index and use location index, fingerprint index, and scaling index to quantify the topology, quality, and quantity of EM nodes, respectively. Besides, three node selection criteria were used: one based on random selection, another using the K-means clustering, and the third involving affinity propagation (AP) clustering. Under each criterion, the overall EM maps were reconstructed using four representative interpolation methods: inverse distance weighting (IDW), modified Shepard’s method (MSM), Kriging algorithm (KGA), and biharmonic spline interpolation (v4). Extensive experiments were conducted to verify the representation model. Experimental results indicate that the node selection criterion based on AP clustering outperforms the K-means criterion, which itself is superior to random selection criterion. This is attributed to the fact that under the same conditions, the AP clustering node selection criterion can better take the quality and the topological relationship of EM nodes into account. Another significant finding is that with good node selection criterion and reconstruction algorithms, only approximately 0.3%–0.5% of the nodes are required to achieve the reconstruction performance that would otherwise require up to 20% or more of the nodes. The above conclusions are of good guidance for the reliable and rapid reconstruction of EM maps via sparse or limited nodes.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-10"},"PeriodicalIF":5.6,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143388583","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：