Pub Date : 2024-07-04DOI: 10.1088/1361-6501/ad5f50
yu liu, jie shen, ruifan ye, shu wang, jia ren, Haipeng Pan
In the pursuit of fabric production efficiency and quality, the application of deep learning for defect detection has become prevalent. Nevertheless, fabric defect detection faces challenges such as low recognition ratio, suboptimal classification performance, detection inefficiency, and high model complexity. To address these issues, an end-to-end semantic segmentation network is proposed employing an efficient encoder-decoder structure, denoted as Feature Pyramid-Deeplab (FP-Deeplab). The improvements involves enhancing the backbone network by improving the mobilenetv3 network for superior performance, a novel Atrous Spatial Pyramid Pooling with Dilated Strip Pooling (ASPP-DSP) module which combines strip pooling, dilated convolution and ASPP, to ensure an expanded receptive field and the capability to gather distant contextual information. Additionally, a Feature Pyramid module (FP module) is proposed to integrate multiscale features at various stages more efficiently. The incorporating of depth-wise separable convolution in FP-Deeplab enables significant parameter and computational cost reduction, catering to online detection requirements. Experimental results showcase the superiority of FP-Deeplab over classical and recent segmentation models. Comparative analysis demonstrates higher segmentation accuracy and reduced parameter quantity. Specifically, compared to the benchmark Deeplabv3+ model with MobileV2 as the backbone, FP-Deeplab achieves a notable increase in segmentation accuracy (F1 score and MIoU) by 4.26% and 5.81%, respectively. Moreover, the model parameters (params) are only one-fifth of the original model, indicating the efficiency and effectiveness of our proposed approach.
{"title":"FP-Deeplab: A Segmentation Model for Fabric Defect Detection","authors":"yu liu, jie shen, ruifan ye, shu wang, jia ren, Haipeng Pan","doi":"10.1088/1361-6501/ad5f50","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5f50","url":null,"abstract":"\u0000 In the pursuit of fabric production efficiency and quality, the application of deep learning for defect detection has become prevalent. Nevertheless, fabric defect detection faces challenges such as low recognition ratio, suboptimal classification performance, detection inefficiency, and high model complexity. To address these issues, an end-to-end semantic segmentation network is proposed employing an efficient encoder-decoder structure, denoted as Feature Pyramid-Deeplab (FP-Deeplab). The improvements involves enhancing the backbone network by improving the mobilenetv3 network for superior performance, a novel Atrous Spatial Pyramid Pooling with Dilated Strip Pooling (ASPP-DSP) module which combines strip pooling, dilated convolution and ASPP, to ensure an expanded receptive field and the capability to gather distant contextual information. Additionally, a Feature Pyramid module (FP module) is proposed to integrate multiscale features at various stages more efficiently. The incorporating of depth-wise separable convolution in FP-Deeplab enables significant parameter and computational cost reduction, catering to online detection requirements. Experimental results showcase the superiority of FP-Deeplab over classical and recent segmentation models. Comparative analysis demonstrates higher segmentation accuracy and reduced parameter quantity. Specifically, compared to the benchmark Deeplabv3+ model with MobileV2 as the backbone, FP-Deeplab achieves a notable increase in segmentation accuracy (F1 score and MIoU) by 4.26% and 5.81%, respectively. Moreover, the model parameters (params) are only one-fifth of the original model, indicating the efficiency and effectiveness of our proposed approach.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141677269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-04DOI: 10.1088/1361-6501/ad5f4f
Rong Rong, C. Shravage, G. Selva Mary, A. John Blesswin, Gayathri M, A. Catherine Esther Karunya, R. Shibani, A. Sambas
Visual Cryptography (VC) has emerged as a vital technique in the information security domain, with the fundamental purpose of securing 2-Dimensional (2D) image content through encryption and facilitating secure communication. While traditional VC has been instrumental in safeguarding data, it often falls short in maintaining image quality and semantic accuracy upon reconstruction. To address these limitations, this research encompasses the development of an Enhanced Semantic Visual Cryptography (ESVC) model, which aims to refine the encryption process while ensuring the semantic integrity of the images. The ESVC model introduces a new approach that merges visual cryptography with artificial intelligence to enhance 2D image encryption and decryption. The novel aspect of this research lies in the integration of AI-driven reinforcement learning to increase the quality of the 2D image by measuring the errors between the original secret image and the reconstructed image. This innovative framework is tailored for the secure transmission of 2D grayscale images, ensuring the preservation of semantic integrity while measuring and minimizing image quality loss. By integrating reinforcement learning algorithms with a measurement of error reduction protocol, the model promises robust encryption capabilities with enhanced resilience against a plethora of cyber threats, thereby elevating the standard for secure image communication. Empirical evaluation of the ESVC model yields promising results, with the Peak Signal-to-Noise Ratio (PSNR) of reconstructed images achieving impressive values between +39 and +42 decibels (dB). These findings underscore the ESVC model's capability to produce high-fidelity decrypted images, significantly surpassing traditional VC methods in both security and image quality. The research findings illuminate the potential of merging AI with visual cryptography to achieve a harmonious balance between computational efficiency and encryption strength, marking a significant advancement in the domain of visual data protection.
{"title":"Enhanced Semantic Visual Cryptography with AI-Driven Error Reduction for Improved two-dimensional Image Quality and Security","authors":"Rong Rong, C. Shravage, G. Selva Mary, A. John Blesswin, Gayathri M, A. Catherine Esther Karunya, R. Shibani, A. Sambas","doi":"10.1088/1361-6501/ad5f4f","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5f4f","url":null,"abstract":"\u0000 Visual Cryptography (VC) has emerged as a vital technique in the information security domain, with the fundamental purpose of securing 2-Dimensional (2D) image content through encryption and facilitating secure communication. While traditional VC has been instrumental in safeguarding data, it often falls short in maintaining image quality and semantic accuracy upon reconstruction. To address these limitations, this research encompasses the development of an Enhanced Semantic Visual Cryptography (ESVC) model, which aims to refine the encryption process while ensuring the semantic integrity of the images. The ESVC model introduces a new approach that merges visual cryptography with artificial intelligence to enhance 2D image encryption and decryption. The novel aspect of this research lies in the integration of AI-driven reinforcement learning to increase the quality of the 2D image by measuring the errors between the original secret image and the reconstructed image. This innovative framework is tailored for the secure transmission of 2D grayscale images, ensuring the preservation of semantic integrity while measuring and minimizing image quality loss. By integrating reinforcement learning algorithms with a measurement of error reduction protocol, the model promises robust encryption capabilities with enhanced resilience against a plethora of cyber threats, thereby elevating the standard for secure image communication. Empirical evaluation of the ESVC model yields promising results, with the Peak Signal-to-Noise Ratio (PSNR) of reconstructed images achieving impressive values between +39 and +42 decibels (dB). These findings underscore the ESVC model's capability to produce high-fidelity decrypted images, significantly surpassing traditional VC methods in both security and image quality. The research findings illuminate the potential of merging AI with visual cryptography to achieve a harmonious balance between computational efficiency and encryption strength, marking a significant advancement in the domain of visual data protection.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141678248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-04DOI: 10.1088/1361-6501/ad5f4e
Renwu Yan, Weiling Zhuang, Ning Yu
The vibration signal associated with the operating process of circuit breakers(CBs) includes a detailed operating status in the formation of the operating mechanism. To effectively extract the characteristic information of vibration effectively for diagnosis and analysis, a new feature extraction method for the CBs operating mechanism is proposed. First, a new denoising method, the improved complete ensemble empirical mode decomposition with adaptive noise-multi-resolution singular value decomposition (ICEEMDAN-MRSVD), is introduced, which can effectively remove the influence of noise on faults. Then, a quantitative method is proposed to extract the characteristic information of the CB, i.e. the variational mode decomposition (VMD)-power spectrum entropy (PSE) is proposed. By using this method, the difference of CB vibration signals in different fault states can be quantified. Through comparative analysis of different recognition models, experiments show that the support vector machine model based on ICEEMDAN-MRSVD noise reduction and VMD-PSE features has a high recognition accuracy of 98.61%, which has high application value.
{"title":"Research on circuit breaker operating mechanism feature extraction method combining ICEEMDAN-MRSVD denoising and VMD-PSE","authors":"Renwu Yan, Weiling Zhuang, Ning Yu","doi":"10.1088/1361-6501/ad5f4e","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5f4e","url":null,"abstract":"\u0000 The vibration signal associated with the operating process of circuit breakers(CBs) includes a detailed operating status in the formation of the operating mechanism. To effectively extract the characteristic information of vibration effectively for diagnosis and analysis, a new feature extraction method for the CBs operating mechanism is proposed. First, a new denoising method, the improved complete ensemble empirical mode decomposition with adaptive noise-multi-resolution singular value decomposition (ICEEMDAN-MRSVD), is introduced, which can effectively remove the influence of noise on faults. Then, a quantitative method is proposed to extract the characteristic information of the CB, i.e. the variational mode decomposition (VMD)-power spectrum entropy (PSE) is proposed. By using this method, the difference of CB vibration signals in different fault states can be quantified. Through comparative analysis of different recognition models, experiments show that the support vector machine model based on ICEEMDAN-MRSVD noise reduction and VMD-PSE features has a high recognition accuracy of 98.61%, which has high application value.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141678098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5eab
Yongyong Hui, Ke Xu, Peng Chen, Xiaomei Zhao
Considering the issue of capturing the local and global contextual information and enhancing the parallel capability of bearing fault diagnosis in variable load and noise environments, a fault diagnosis method of rolling bearing based on PE-DCM and ViT is proposed. Firstly, the one-dimensional vibration signal is converted into a two-dimensional time-frequency diagram by continuous wavelet transform in the data processing module, and the model can understand the characteristics of the vibration signal more comprehensively. Secondly, a pyramid exponential expansion convolution module is established to extract the local features of fault information. Then, the global features of the fault information are learnt through the ViT (Vision Transformer) network, and the adaptive multi-attention is used to dynamically adjust the attention weights according to the features of the input data so as to inhibit noise or unimportant information. Finally, the experimental verification is carried out by using Case Western Reserve University and self-made MFS-bearing data set. The experimental results show that the method can better reflect the powerful image classification ability of the ViT network and has better noise resistance and generalization compared with other fault diagnosis methods.
考虑到在变载荷和噪声环境下捕捉局部和全局上下文信息、提高轴承故障诊断并行能力的问题,提出了一种基于 PE-DCM 和 ViT 的滚动轴承故障诊断方法。首先,在数据处理模块中通过连续小波变换将一维振动信号转换为二维时频图,该模型能更全面地了解振动信号的特征。其次,建立金字塔指数膨胀卷积模块,提取故障信息的局部特征。然后,通过 ViT(Vision Transformer)网络学习故障信息的全局特征,并利用自适应多注意功能根据输入数据的特征动态调整注意权重,以抑制噪声或不重要的信息。最后,利用凯斯西储大学和自制的 MFS 负载数据集进行了实验验证。实验结果表明,与其他故障诊断方法相比,该方法能更好地体现 ViT 网络强大的图像分类能力,并具有更好的抗噪性和泛化能力。
{"title":"Rolling bearing fault diagnosis method based on PE-DCM and ViT","authors":"Yongyong Hui, Ke Xu, Peng Chen, Xiaomei Zhao","doi":"10.1088/1361-6501/ad5eab","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5eab","url":null,"abstract":"\u0000 Considering the issue of capturing the local and global contextual information and enhancing the parallel capability of bearing fault diagnosis in variable load and noise environments, a fault diagnosis method of rolling bearing based on PE-DCM and ViT is proposed. Firstly, the one-dimensional vibration signal is converted into a two-dimensional time-frequency diagram by continuous wavelet transform in the data processing module, and the model can understand the characteristics of the vibration signal more comprehensively. Secondly, a pyramid exponential expansion convolution module is established to extract the local features of fault information. Then, the global features of the fault information are learnt through the ViT (Vision Transformer) network, and the adaptive multi-attention is used to dynamically adjust the attention weights according to the features of the input data so as to inhibit noise or unimportant information. Finally, the experimental verification is carried out by using Case Western Reserve University and self-made MFS-bearing data set. The experimental results show that the method can better reflect the powerful image classification ability of the ViT network and has better noise resistance and generalization compared with other fault diagnosis methods.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141681927","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The production process of lithium batteries is intricate, involving the coordination of various types of equipment.The sta-bility and precision of double roller press equipment directly affect product performance. With the increasing global de-mand for green energy, the application of lithium batteries in electric vehicles and energy storage systems is expanding, which imposes higher requirements on the stability and quality of lithium battery production. It is an important topic to address the challenges brought about by the gradual intelligentization of double roller presses, such as the complexifica-tion of control systems and the diversification of fault reasons. This paper proposes an enhanced ensemble learning model-based fault diagnosis strategy for lithium battery double roller press equipment. Firstly, the K-nearest neighbors (KNN) algorithm is employed to handle missing data, combined with normalization and standardization methods to improve fea-ture processing, thereby enhancing data quality. Secondly, the Maximum Information Coefficient (MIC) algorithm is utilized to select features highly correlated with fault labels, combined with the Recursive Feature Elimination with Cross-Validation (RFECV) to further optimize feature selection, creating an optimal feature subset. Finally, a RXS-XGBoost model is constructed through the Stacking ensemble learning method, selecting Random Forest (RF), XGBoost, and Sup-port Vector Machines (SVM) as base learners, with XGBoost as the meta-learner. This ensemble approach aims to lever-age the complementary advantages of different algorithms, enhancing the accuracy and robustness of fault diagnosis. The experimental results demonstrate that this improved ensemble learning diagnostic strategy achieves an accuracy rate of up to 99.05%, which is significantly better than other fault diagnosis strategies. It not only effectively reduces the model's training complexity and the risk of overfitting but also significantly enhances the efficiency and precision of fault diagno-sis for lithium battery double roller press equipment.
{"title":"An Improved Ensemble Learning Model-Based Strategy for Fault Diagnosis of Lithium Battery Double Roller Press Equipment","authors":"Yanjun Xiao, Weihan Song, Shanshan Yin, Feng Wan, Weiling Liu, Nannan Zhang","doi":"10.1088/1361-6501/ad5ea0","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5ea0","url":null,"abstract":"\u0000 The production process of lithium batteries is intricate, involving the coordination of various types of equipment.The sta-bility and precision of double roller press equipment directly affect product performance. With the increasing global de-mand for green energy, the application of lithium batteries in electric vehicles and energy storage systems is expanding, which imposes higher requirements on the stability and quality of lithium battery production. It is an important topic to address the challenges brought about by the gradual intelligentization of double roller presses, such as the complexifica-tion of control systems and the diversification of fault reasons. This paper proposes an enhanced ensemble learning model-based fault diagnosis strategy for lithium battery double roller press equipment. Firstly, the K-nearest neighbors (KNN) algorithm is employed to handle missing data, combined with normalization and standardization methods to improve fea-ture processing, thereby enhancing data quality. Secondly, the Maximum Information Coefficient (MIC) algorithm is utilized to select features highly correlated with fault labels, combined with the Recursive Feature Elimination with Cross-Validation (RFECV) to further optimize feature selection, creating an optimal feature subset. Finally, a RXS-XGBoost model is constructed through the Stacking ensemble learning method, selecting Random Forest (RF), XGBoost, and Sup-port Vector Machines (SVM) as base learners, with XGBoost as the meta-learner. This ensemble approach aims to lever-age the complementary advantages of different algorithms, enhancing the accuracy and robustness of fault diagnosis. The experimental results demonstrate that this improved ensemble learning diagnostic strategy achieves an accuracy rate of up to 99.05%, which is significantly better than other fault diagnosis strategies. It not only effectively reduces the model's training complexity and the risk of overfitting but also significantly enhances the efficiency and precision of fault diagno-sis for lithium battery double roller press equipment.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141682128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5ea5
L. Cao, Wentao Guo, Binyan He, Weihong Li, Xufeng Huang, Y. Zhang, Wang Cai, Qi Zhou
The small changes in process parameters have significant influences on the stability of laser powder bed fusion (LPBF). Therefore, monitoring the small changes in process parameters is particularly important. This paper proposed a machine learning (ML)-based multi-sensor fusion approach to monitor the LPBF processing state by combining photodiode, acoustic, and visual signals. In order to extract the motion features of the melt pool more accurately and describe its transient changes, an ellipse adjustment algorithm is proposed to segment the melt pool images, eliminating the interference of spatters. The motion features combined with preprocessed acoustic signals and photodiode signals to identify melting states during small changes in process parameters. The proposed ML-based multi-sensor fusion approach achieves impressive prediction accuracies of 99.9% for identifying the fluctuations in the process parameters. The results demonstrate that the proposed method can accurately identify small changes in process parameters, which is of great significance for improving the process stability and providing reliable guidance in subsequent work.
工艺参数的微小变化会对激光粉末床熔融(LPBF)的稳定性产生重大影响。因此,监测工艺参数的微小变化尤为重要。本文提出了一种基于机器学习(ML)的多传感器融合方法,通过结合光电二极管、声学和视觉信号来监测 LPBF 的加工状态。为了更准确地提取熔池的运动特征并描述其瞬态变化,提出了一种椭圆调整算法来分割熔池图像,消除了飞溅物的干扰。运动特征与预处理的声学信号和光电二极管信号相结合,可识别工艺参数微小变化时的熔化状态。所提出的基于 ML 的多传感器融合方法在识别工艺参数波动方面的预测准确率高达 99.9%,令人印象深刻。结果表明,所提出的方法可以准确识别工艺参数的微小变化,这对于提高工艺稳定性和为后续工作提供可靠指导具有重要意义。
{"title":"In-situ monitoring of the small changes in process parameters with multi-sensor fusion during LPBF","authors":"L. Cao, Wentao Guo, Binyan He, Weihong Li, Xufeng Huang, Y. Zhang, Wang Cai, Qi Zhou","doi":"10.1088/1361-6501/ad5ea5","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5ea5","url":null,"abstract":"\u0000 The small changes in process parameters have significant influences on the stability of laser powder bed fusion (LPBF). Therefore, monitoring the small changes in process parameters is particularly important. This paper proposed a machine learning (ML)-based multi-sensor fusion approach to monitor the LPBF processing state by combining photodiode, acoustic, and visual signals. In order to extract the motion features of the melt pool more accurately and describe its transient changes, an ellipse adjustment algorithm is proposed to segment the melt pool images, eliminating the interference of spatters. The motion features combined with preprocessed acoustic signals and photodiode signals to identify melting states during small changes in process parameters. The proposed ML-based multi-sensor fusion approach achieves impressive prediction accuracies of 99.9% for identifying the fluctuations in the process parameters. The results demonstrate that the proposed method can accurately identify small changes in process parameters, which is of great significance for improving the process stability and providing reliable guidance in subsequent work.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141680612","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5ea6
Wei Xie, Weiming Liu, Y. Dai
Detecting vehicles from a surveillance viewpoint is essential, as it has wide applications in community security and traffic control. However, existing methods completely overlook the high memory access costs (MAC) and low degree of parallelism inherent in multi-branch topologies, resulting in significant latency during inference. Additionally, existing methods share the same positive/negative sample set between the classification and localization branches, leading to sample misalignment, and rely solely on intersection-over-union (IoU) for sample assignment, thereby causing a decrease in detection performance. To tackle these issues, this paper introduces a two-stage re-parameterization and sample disentanglement network (TRSD-Net). It is based on two-stage depthwise to pointwise re-parameterization (RepTDP) and task-aligned sample disentanglement (TSD). RepTDP employs structural re-parameterization to decouple the multi-branch topology during training and the plain architecture during inference, thus achieving low latency. By employing different sample assignment strategies, TSD can adaptively select the most suitable positive/negative sample sets for classification and localization tasks, thereby enhancing detection performance. Additionally, TSD considers three important factors influencing sample assignment. TRSD-Net is evaluated on both the UA-DETRAC and COCO datasets. On the UA-DETRAC dataset, compared to state-of-the-art (SOTA) methods, TRSD-Net improves the detection accuracy from 58.8% to 59.7%. Additionally, it reduces the parameter count by 87%, the computational complexity by 85%, and the latency by 39%, while increasing the detection speed by 65%. Similar performance improvement trends are observed on the COCO dataset.
{"title":"Two-Stage Re-Parameterization and Sample Disentanglement Network for Surveillance Vehicle Detection","authors":"Wei Xie, Weiming Liu, Y. Dai","doi":"10.1088/1361-6501/ad5ea6","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5ea6","url":null,"abstract":"\u0000 Detecting vehicles from a surveillance viewpoint is essential, as it has wide applications in community security and traffic control. However, existing methods completely overlook the high memory access costs (MAC) and low degree of parallelism inherent in multi-branch topologies, resulting in significant latency during inference. Additionally, existing methods share the same positive/negative sample set between the classification and localization branches, leading to sample misalignment, and rely solely on intersection-over-union (IoU) for sample assignment, thereby causing a decrease in detection performance. To tackle these issues, this paper introduces a two-stage re-parameterization and sample disentanglement network (TRSD-Net). It is based on two-stage depthwise to pointwise re-parameterization (RepTDP) and task-aligned sample disentanglement (TSD). RepTDP employs structural re-parameterization to decouple the multi-branch topology during training and the plain architecture during inference, thus achieving low latency. By employing different sample assignment strategies, TSD can adaptively select the most suitable positive/negative sample sets for classification and localization tasks, thereby enhancing detection performance. Additionally, TSD considers three important factors influencing sample assignment. TRSD-Net is evaluated on both the UA-DETRAC and COCO datasets. On the UA-DETRAC dataset, compared to state-of-the-art (SOTA) methods, TRSD-Net improves the detection accuracy from 58.8% to 59.7%. Additionally, it reduces the parameter count by 87%, the computational complexity by 85%, and the latency by 39%, while increasing the detection speed by 65%. Similar performance improvement trends are observed on the COCO dataset.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141682637","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5eaa
Eberhard Manske, Thomas Fröhlich, Thomas Kissinger
The 60th Ilmenau Scientific Colloquium was held from 4th to 8th September 2023 at the Technische Universität Ilmenau in Germany. Organized by the Faculty of Mechanical Engineering under the title 'Engineering for a Changing World', it was intended to focus on the many challenges facing modern mechanical engineering.
{"title":"Precision Measurement and Engineering at the 60th Ilmenau Scientific Colloquium","authors":"Eberhard Manske, Thomas Fröhlich, Thomas Kissinger","doi":"10.1088/1361-6501/ad5eaa","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5eaa","url":null,"abstract":"\u0000 The 60th Ilmenau Scientific Colloquium was held from 4th to 8th September 2023 at the Technische Universität Ilmenau in Germany. Organized by the Faculty of Mechanical Engineering under the title 'Engineering for a Changing World', it was intended to focus on the many challenges facing modern mechanical engineering.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141681662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5ea3
Huan Yang, Xin Li, Yuan Du, Ce Jing, Guolin Liu, Kai Zhang, Xiaoyu Huang
In the field of structural health monitoring (SHM), a loosely coupled (LC) Kalman filtering algorithm that accounts for baseline drift errors is commonly used to integrate GNSS data with accelerometer data. In the LC algorithm, the baseline drift errors are considered unknown parameters that need to be estimated. In scenario of continuous float solutions, the estimation of baseline drift error is often inaccurate, leading to the divergence of monitoring results. Theoretically, as a type of motion sensor, accelerometers are expected to qualitatively determine the priori state of bridges, whether dynamic or static. Utilizing the inherent characteristics of accelerometers and the principle of zero-velocity detection in integrated navigation, we originally propose a bridge static state detection (SSD) method based on low-cost accelerometer, and introduces this prior SSD information as a constraint in GNSS/accelerometer LC algorithm, called SSD-LC bridge monitoring algorithm. Through a simulation platform and real-world bridge monitored tests, the effectiveness of our proposed SSD method has been verified. Furthermore, our proposed SSD-LC bridge monitoring algorithm can effectively mitigate the divergence problem in baseline drift estimation that occurs with continuous GNSS float solutions in traditional algorithms, which can effectively avoid misjudgments and false alarms in bridge monitoring during GNSS anomalies.
{"title":"Accelerometer static state detection (SSD)-assisted GNSS/accelerometer bridge monitoring algorithm","authors":"Huan Yang, Xin Li, Yuan Du, Ce Jing, Guolin Liu, Kai Zhang, Xiaoyu Huang","doi":"10.1088/1361-6501/ad5ea3","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5ea3","url":null,"abstract":"\u0000 In the field of structural health monitoring (SHM), a loosely coupled (LC) Kalman filtering algorithm that accounts for baseline drift errors is commonly used to integrate GNSS data with accelerometer data. In the LC algorithm, the baseline drift errors are considered unknown parameters that need to be estimated. In scenario of continuous float solutions, the estimation of baseline drift error is often inaccurate, leading to the divergence of monitoring results. Theoretically, as a type of motion sensor, accelerometers are expected to qualitatively determine the priori state of bridges, whether dynamic or static. Utilizing the inherent characteristics of accelerometers and the principle of zero-velocity detection in integrated navigation, we originally propose a bridge static state detection (SSD) method based on low-cost accelerometer, and introduces this prior SSD information as a constraint in GNSS/accelerometer LC algorithm, called SSD-LC bridge monitoring algorithm. Through a simulation platform and real-world bridge monitored tests, the effectiveness of our proposed SSD method has been verified. Furthermore, our proposed SSD-LC bridge monitoring algorithm can effectively mitigate the divergence problem in baseline drift estimation that occurs with continuous GNSS float solutions in traditional algorithms, which can effectively avoid misjudgments and false alarms in bridge monitoring during GNSS anomalies.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141681693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-07-03DOI: 10.1088/1361-6501/ad5ea7
Huadong Yang, Rongxin Song, Geng Ma, Jianhua Wang
In the field of metallic additive manufacturing, Selective Laser Melting (SLM) has become a predominant technology due to its advantages of short production cycles, high precision, and low cost. It is frequently employed in the production of complex parts. This paper proposes the use of a scanning laser line source, in conjunction with the singular value decomposition method, to reconstruct phase space and identify surface cracks in SLM specimens. The scanning laser line source addresses the limitations of a single line source, which is often unable to accurately detect tiny cracks. By comparing experimental and simulation data, the results demonstrate that the scanning laser line source can effectively compensate for some of the detection deficiencies of a single line source.
{"title":"Surface Crack Detection on Selective Laser Melting Printed Inconel 718 Using a Laser Generated Ultrasound Technique and Phase Space Reconstruction","authors":"Huadong Yang, Rongxin Song, Geng Ma, Jianhua Wang","doi":"10.1088/1361-6501/ad5ea7","DOIUrl":"https://doi.org/10.1088/1361-6501/ad5ea7","url":null,"abstract":"\u0000 In the field of metallic additive manufacturing, Selective Laser Melting (SLM) has become a predominant technology due to its advantages of short production cycles, high precision, and low cost. It is frequently employed in the production of complex parts. This paper proposes the use of a scanning laser line source, in conjunction with the singular value decomposition method, to reconstruct phase space and identify surface cracks in SLM specimens. The scanning laser line source addresses the limitations of a single line source, which is often unable to accurately detect tiny cracks. By comparing experimental and simulation data, the results demonstrate that the scanning laser line source can effectively compensate for some of the detection deficiencies of a single line source.","PeriodicalId":18526,"journal":{"name":"Measurement Science and Technology","volume":null,"pages":null},"PeriodicalIF":2.7,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141682597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}