Computers in Industry最新文献_第2页

Measure2Shape: A novel footwear customisation framework utilising 3D shape estimation from anthropometric measurements with an orthosis case study

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-05 DOI: 10.1016/j.compind.2025.104257

Zhaohua Zhu , Wenxuan Ji , Yadie Yang , Sio-Kei Im , Jie Zhang

To address the limitations of relying on expensive 3D scanners for obtaining foot data in footwear customisation, this paper introduces a novel framework, Measure2Shape, which estimates 3D foot shapes using anthropometric measurement data. To achieve this, we established a large-scale 3D foot dataset with measurement data and created statistical shape models (SSMs) to represent the range of foot variations. We then proposed efficient forward- and backward-search algorithms to accurately determine the regression matrix, which connects the optimal combination of 3D measurements to the SSM coefficients of the 3D foot shape. Compared to existing 3D foot model estimation methods, our approach achieves high-precision 3D foot shape predictions using fewer dimensional measurements, with the optimal number being 6 and an average prediction error of 2.49 (±0.75) mm. Additionally, orthosis designed based on the predicted 3D foot model effectively reduce both static and dynamic peak plantar pressures, validating the reliability of our model. More importantly, the proposed regression search method can be extended to 3D estimations for other body regions, offering a wide range of customisation solutions beyond footwear. In the future, we will further expand the dataset to build a more robust 3D foot prediction model. Our project will be publicly available at: https://github.com/Easy-Shu/Measure2Shape.

{"title":"Measure2Shape: A novel footwear customisation framework utilising 3D shape estimation from anthropometric measurements with an orthosis case study","authors":"Zhaohua Zhu , Wenxuan Ji , Yadie Yang , Sio-Kei Im , Jie Zhang","doi":"10.1016/j.compind.2025.104257","DOIUrl":"10.1016/j.compind.2025.104257","url":null,"abstract":"<div><div>To address the limitations of relying on expensive 3D scanners for obtaining foot data in footwear customisation, this paper introduces a novel framework, Measure2Shape, which estimates 3D foot shapes using anthropometric measurement data. To achieve this, we established a large-scale 3D foot dataset with measurement data and created statistical shape models (SSMs) to represent the range of foot variations. We then proposed efficient forward- and backward-search algorithms to accurately determine the regression matrix, which connects the optimal combination of 3D measurements to the SSM coefficients of the 3D foot shape. Compared to existing 3D foot model estimation methods, our approach achieves high-precision 3D foot shape predictions using fewer dimensional measurements, with the optimal number being 6 and an average prediction error of 2.49 (±0.75) mm. Additionally, orthosis designed based on the predicted 3D foot model effectively reduce both static and dynamic peak plantar pressures, validating the reliability of our model. More importantly, the proposed regression search method can be extended to 3D estimations for other body regions, offering a wide range of customisation solutions beyond footwear. In the future, we will further expand the dataset to build a more robust 3D foot prediction model. Our project will be publicly available at: <span><span>https://github.com/Easy-Shu/Measure2Shape</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"167 ","pages":"Article 104257"},"PeriodicalIF":8.2,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143125022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An immersive spatially consistent multi-modal augmented virtuality human-machine interface for telerobotic systems

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-05 DOI: 10.1016/j.compind.2025.104260

Rebecca Schwenk, Shana Smith

This study presents an immersive augmented virtuality (AV)-based human-machine interface (HMI) designed to enhance telepresence and operator performance in telerobotic systems. Traditional telerobotic systems often face limitations such as 2D representations of 3D environments, restricted fields of view, and reduced depth perception, all of which hinder operator effectiveness. Although extended reality and various augmentation technologies have been employed to create more intuitive teleoperations, prior research has largely overlooked the integration of spatially consistent video streams from remote sites, which significantly increases operators' mental workload. As a result, these systems struggled to manage dynamic changes at the remote site and lacked sufficient environmental context and an unlimited field of view for operators. This study addresses these limitations by augmenting the virtual replica of the remote environment with a real-time, spatially consistent video stream within the AV-based HMI, enabling operators to better understand dynamic changes at the remote site and enhancing both situational awareness and control precision during teleoperations. Additionally, 3D point clouds and haptic feedback are integrated to create a multi-modal interface that further improves operator perception and interaction with the remote environment. A user study comparing the immersive AV-based HMI with a multi-monocular HMI demonstrated significant improvements in task workload, system usability, spatial presence, and task completion times. Participant feedback further confirmed the system’s ability to improve operator performance.

{"title":"An immersive spatially consistent multi-modal augmented virtuality human-machine interface for telerobotic systems","authors":"Rebecca Schwenk, Shana Smith","doi":"10.1016/j.compind.2025.104260","DOIUrl":"10.1016/j.compind.2025.104260","url":null,"abstract":"<div><div>This study presents an immersive augmented virtuality (AV)-based human-machine interface (HMI) designed to enhance telepresence and operator performance in telerobotic systems. Traditional telerobotic systems often face limitations such as 2D representations of 3D environments, restricted fields of view, and reduced depth perception, all of which hinder operator effectiveness. Although extended reality and various augmentation technologies have been employed to create more intuitive teleoperations, prior research has largely overlooked the integration of spatially consistent video streams from remote sites, which significantly increases operators' mental workload. As a result, these systems struggled to manage dynamic changes at the remote site and lacked sufficient environmental context and an unlimited field of view for operators. This study addresses these limitations by augmenting the virtual replica of the remote environment with a real-time, spatially consistent video stream within the AV-based HMI, enabling operators to better understand dynamic changes at the remote site and enhancing both situational awareness and control precision during teleoperations. Additionally, 3D point clouds and haptic feedback are integrated to create a multi-modal interface that further improves operator perception and interaction with the remote environment. A user study comparing the immersive AV-based HMI with a multi-monocular HMI demonstrated significant improvements in task workload, system usability, spatial presence, and task completion times. Participant feedback further confirmed the system’s ability to improve operator performance.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"167 ","pages":"Article 104260"},"PeriodicalIF":8.2,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143125282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Evaluating unsignalized crosswalk safety in the age of autonomous vehicles

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-03 DOI: 10.1016/j.compind.2025.104259

Andrea Avignone , Marco Bassani , Beatrice Borgogno , Brunella Caroleo , Silvia Chiusano , Federico Princiotto

As autonomous vehicles are poised to enter public roadways, a major concern is their interaction with pedestrians. It requires attention and ability for pedestrians to interact correctly and for autonomous vehicles to detect pedestrians hence avoiding collisions. We propose a complete pipeline to collect, process and elaborate video data to quantitatively assess the possible occurrence of conflicts. It integrates computer vision techniques and a conflict detection system to evaluate these interactions by rigorously implementing the theoretical formulation of two primary metrics: Time-to-Collision (TTC) for the pre-event phase and Post Encroachment Time (PET) for the post-event phase. This study is conducted in a real-world setting with mixed traffic conditions to analyse the differences in pedestrian interactions with both human-operated and autonomous vehicles during daytime. The computation of conflict measures allowed us to identify possible conflicts and assess the safety at an unsignalized crossing, in which pedestrians are exposed to more risky conflicts. The results obtained show a higher incidence of more severe conflicts for interactions between pedestrians and human-operated vehicles, which highlights the caution taken in programming the autonomous vehicle.

{"title":"Evaluating unsignalized crosswalk safety in the age of autonomous vehicles","authors":"Andrea Avignone , Marco Bassani , Beatrice Borgogno , Brunella Caroleo , Silvia Chiusano , Federico Princiotto","doi":"10.1016/j.compind.2025.104259","DOIUrl":"10.1016/j.compind.2025.104259","url":null,"abstract":"<div><div>As autonomous vehicles are poised to enter public roadways, a major concern is their interaction with pedestrians. It requires attention and ability for pedestrians to interact correctly and for autonomous vehicles to detect pedestrians hence avoiding collisions. We propose a complete pipeline to collect, process and elaborate video data to quantitatively assess the possible occurrence of conflicts. It integrates computer vision techniques and a conflict detection system to evaluate these interactions by rigorously implementing the theoretical formulation of two primary metrics: Time-to-Collision (TTC) for the pre-event phase and Post Encroachment Time (PET) for the post-event phase. This study is conducted in a real-world setting with mixed traffic conditions to analyse the differences in pedestrian interactions with both human-operated and autonomous vehicles during daytime. The computation of conflict measures allowed us to identify possible conflicts and assess the safety at an unsignalized crossing, in which pedestrians are exposed to more risky conflicts. The results obtained show a higher incidence of more severe conflicts for interactions between pedestrians and human-operated vehicles, which highlights the caution taken in programming the autonomous vehicle.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"167 ","pages":"Article 104259"},"PeriodicalIF":8.2,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143125024","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Intelligent prediction and soft-sensing of comprehensive production indicators for iron ore sintering: A review 铁矿石烧结综合生产指标的智能预测与软测量研究进展

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104215

Sheng Du , Xian Ma , Haipeng Fan , Jie Hu , Weihua Cao , Min Wu , Witold Pedrycz

Iron ore sintering is a critical process in iron and steel production, with a substantial impact on overall energy consumption and the emission of various environmental pollutants. Enhancing the efficiency of this process is crucial for achieving sustainability in the iron and steel industry. Accurate prediction and real-time monitoring of comprehensive production indicators are essential for optimizing production and improving energy efficiency. This paper provides a systematic review of intelligent prediction and soft-sensing techniques applied to the iron ore sintering process. It details the mechanisms and operational principles of these technologies, with a focus on key indicators such as quality, thermal state, yield, and energy consumption. This paper explores the current state-of-the-art in four prediction methodologies: mechanism analysis-based methods, data feature analysis-based methods, multi-model fusion-based methods, and operating mode recognition-based methods. Finally, the challenges to the current comprehensive production indicator prediction of the sintering process are pointed out, including the difficulty of dealing with the changing operating mode, the incomplete analysis of image features, and the insufficient consideration of the differences in data distribution. In the future, operating mode recognition approaches, deep learning approaches, transfer learning approaches, and computer vision techniques will have a broad prospect in the comprehensive production indicator prediction of the sintering process.

铁矿石烧结是钢铁生产中的关键工序，对整体能源消耗和各种环境污染物的排放有重大影响。提高这一过程的效率对于实现钢铁工业的可持续性至关重要。对综合生产指标的准确预测和实时监控是优化生产和提高能源效率的必要条件。本文系统地综述了智能预测和软测量技术在铁矿石烧结过程中的应用。它详细介绍了这些技术的机制和操作原理，重点介绍了质量、热状态、产量和能耗等关键指标。本文探讨了基于机理分析的预测方法、基于数据特征分析的预测方法、基于多模型融合的预测方法和基于运行模式识别的预测方法。最后指出了当前烧结工艺综合生产指标预测面临的挑战，包括难以应对操作模式的变化、对图像特征分析不完整、对数据分布差异考虑不足等。未来，运行模式识别方法、深度学习方法、迁移学习方法和计算机视觉技术在烧结过程综合生产指标预测中具有广阔的前景。

{"title":"Intelligent prediction and soft-sensing of comprehensive production indicators for iron ore sintering: A review","authors":"Sheng Du , Xian Ma , Haipeng Fan , Jie Hu , Weihua Cao , Min Wu , Witold Pedrycz","doi":"10.1016/j.compind.2024.104215","DOIUrl":"10.1016/j.compind.2024.104215","url":null,"abstract":"<div><div>Iron ore sintering is a critical process in iron and steel production, with a substantial impact on overall energy consumption and the emission of various environmental pollutants. Enhancing the efficiency of this process is crucial for achieving sustainability in the iron and steel industry. Accurate prediction and real-time monitoring of comprehensive production indicators are essential for optimizing production and improving energy efficiency. This paper provides a systematic review of intelligent prediction and soft-sensing techniques applied to the iron ore sintering process. It details the mechanisms and operational principles of these technologies, with a focus on key indicators such as quality, thermal state, yield, and energy consumption. This paper explores the current state-of-the-art in four prediction methodologies: mechanism analysis-based methods, data feature analysis-based methods, multi-model fusion-based methods, and operating mode recognition-based methods. Finally, the challenges to the current comprehensive production indicator prediction of the sintering process are pointed out, including the difficulty of dealing with the changing operating mode, the incomplete analysis of image features, and the insufficient consideration of the differences in data distribution. In the future, operating mode recognition approaches, deep learning approaches, transfer learning approaches, and computer vision techniques will have a broad prospect in the comprehensive production indicator prediction of the sintering process.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104215"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142804460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Meta-task interpolation-based data augmentation for imbalanced health status recognition of complex equipment 基于元任务插值的复杂设备不平衡健康状态识别数据增强

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104226

Jinyuan Li, Wenqing Wan, Yong Feng, Jinglong Chen

In the research of health status detection technology for complex equipment such as liquid rocket engines, the extreme working environment hinders the widespread conduct of fault experimental simulations, leading to data scarcity and imbalance. Consequently, the performance of intelligent models deteriorates rapidly with direct training. To address this issue, this paper proposes a meta-task feature space interpolation network model. Firstly, the model uses an encoder to map randomly selected task pairs to a more discriminative feature space, and then interpolates corresponding features and labels within this latent feature space to generate additional tasks, increasing the distribution density of tasks and alleviating the problem of insufficient training tasks. Furthermore, the model leverages self-distillation to improve the learning of label information. By integrating soft labels with supervised labels, it captures the hidden category information of newly interpolated tasks, thereby reducing the impact of class imbalance on model performance. The effectiveness of the proposed method is validated through a series of experiments conducted across three different scenarios. The results demonstrate that the proposed method achieves an average accuracy of 97.91% on the turbopump bearing dataset, which is a significant improvement over the comparative methods.

在液体火箭发动机等复杂设备的健康状态检测技术研究中，极端的工作环境阻碍了故障实验模拟的广泛开展，导致数据稀缺和不平衡。因此，直接训练会导致智能模型的性能迅速下降。为了解决这一问题，本文提出了一种元任务特征空间插值网络模型。该模型首先使用编码器将随机选择的任务对映射到更具判别性的特征空间，然后在该潜在特征空间内插值相应的特征和标签生成额外的任务，从而增加任务的分布密度，缓解训练任务不足的问题。此外，该模型利用自蒸馏来改进标签信息的学习。通过将软标签与监督标签相结合，捕获新插入任务的隐藏类别信息，从而减少类不平衡对模型性能的影响。通过在三种不同场景下进行的一系列实验，验证了所提出方法的有效性。结果表明，该方法在涡轮泵轴承数据集上的平均准确率达到97.91%，与其他对比方法相比有显著提高。

{"title":"Meta-task interpolation-based data augmentation for imbalanced health status recognition of complex equipment","authors":"Jinyuan Li, Wenqing Wan, Yong Feng, Jinglong Chen","doi":"10.1016/j.compind.2024.104226","DOIUrl":"10.1016/j.compind.2024.104226","url":null,"abstract":"<div><div>In the research of health status detection technology for complex equipment such as liquid rocket engines, the extreme working environment hinders the widespread conduct of fault experimental simulations, leading to data scarcity and imbalance. Consequently, the performance of intelligent models deteriorates rapidly with direct training. To address this issue, this paper proposes a meta-task feature space interpolation network model. Firstly, the model uses an encoder to map randomly selected task pairs to a more discriminative feature space, and then interpolates corresponding features and labels within this latent feature space to generate additional tasks, increasing the distribution density of tasks and alleviating the problem of insufficient training tasks. Furthermore, the model leverages self-distillation to improve the learning of label information. By integrating soft labels with supervised labels, it captures the hidden category information of newly interpolated tasks, thereby reducing the impact of class imbalance on model performance. The effectiveness of the proposed method is validated through a series of experiments conducted across three different scenarios. The results demonstrate that the proposed method achieves an average accuracy of 97.91% on the turbopump bearing dataset, which is a significant improvement over the comparative methods.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104226"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142884279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Integration of industry 4.0 technologies for agri-food supply chain resilience 整合工业4.0技术提升农业食品供应链弹性

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104225

Rohit Sharma , Balan Sundarakani , Ioannis Manikas

The agri-food supply chain (AFSC) operations are becoming challenging due to globalization, constantly shifting consumer demands, and intensive disruptions leading to inefficient production and distribution of safe and high-quality food. Technological advancements are the most promising ways to ensure firms’ survival and supply chains. To enhance the resilience of AFSCs, the present study aims to identify and model the challenges associated with AFSC operations in the context of the United Arab Emirates (UAE) food processing industry. An integrated methodology using the Grey Influence Analysis (GINA) and Fuzzy Linguistic Quantifier Ordered Weighted Aggregation (FLQOWA) methodology is applied to analyze resilience enablers and assess industry 4.0 technologies (I4Ts) that can enhance resilience in AFSCs. The GINA technique helps identify the most influential resilience enablers, and the FLQOWA helps assess and prioritize I4Ts to enhance resilient enablers. The findings reveal that out of thirteen sub-enablers, four are the most influential resilient enablers, viz., real-time information sharing, enhanced product traceability, improved risk management, and planning and network design; and out of ten I4Ts, three are the most influential technologies viz., big data analytics, Internet of things, and cloud computing can further enhance resilience enablers. The findings from the study can help AFSC organizations and the government formulate appropriate strategies based on the integrated matrix developed by selecting the best combination of technologies for strengthening the required resilient enablers among the AFSC stakeholders.

由于全球化、消费者需求的不断变化以及导致安全优质食品生产和分销效率低下的密集干扰，农业食品供应链（AFSC）的运营正变得越来越具有挑战性。技术进步是确保企业生存和供应链的最有前途的方法。为了提高农产品加工企业的应变能力，本研究旨在以阿拉伯联合酋长国（UAE）食品加工业为背景，识别与农产品加工企业运营相关的挑战并建立相关模型。本研究采用灰色影响分析（GINA）和模糊语言量化有序加权聚合（FLQOWA）综合方法来分析抗灾能力增强因素，并评估可提高阿联酋食品加工业抗灾能力的工业 4.0 技术（I4T）。GINA 技术有助于确定最具影响力的复原力增强因素，而 FLQOWA 则有助于评估和优先考虑工业 4T 以增强复原力增强因素。研究结果表明，在 13 个子使能因素中，有 4 个是最具影响力的复原力使能因素，即实时信息共享、增强产品可追溯性、改进风险管理以及规划和网络设计；在 10 个 I4T 中，有 3 个是最具影响力的技术，即大数据分析、物联网和云计算可进一步增强复原力使能因素。这项研究的结果可以帮助阿富汗国家安全部队组织和政府根据综合矩阵制定适当的战略，选择最佳的技术组合来加强阿富汗国家安全部队利益攸关方所需的抗灾能力。

{"title":"Integration of industry 4.0 technologies for agri-food supply chain resilience","authors":"Rohit Sharma , Balan Sundarakani , Ioannis Manikas","doi":"10.1016/j.compind.2024.104225","DOIUrl":"10.1016/j.compind.2024.104225","url":null,"abstract":"<div><div>The agri-food supply chain (AFSC) operations are becoming challenging due to globalization, constantly shifting consumer demands, and intensive disruptions leading to inefficient production and distribution of safe and high-quality food. Technological advancements are the most promising ways to ensure firms’ survival and supply chains. To enhance the resilience of AFSCs, the present study aims to identify and model the challenges associated with AFSC operations in the context of the United Arab Emirates (UAE) food processing industry. An integrated methodology using the Grey Influence Analysis (GINA) and Fuzzy Linguistic Quantifier Ordered Weighted Aggregation (FLQOWA) methodology is applied to analyze resilience enablers and assess industry 4.0 technologies (I4Ts) that can enhance resilience in AFSCs. The GINA technique helps identify the most influential resilience enablers, and the FLQOWA helps assess and prioritize I4Ts to enhance resilient enablers. The findings reveal that out of thirteen sub-enablers, four are the most influential resilient enablers, viz., real-time information sharing, enhanced product traceability, improved risk management, and planning and network design; and out of ten I4Ts, three are the most influential technologies viz., big data analytics, Internet of things, and cloud computing can further enhance resilience enablers. The findings from the study can help AFSC organizations and the government formulate appropriate strategies based on the integrated matrix developed by selecting the best combination of technologies for strengthening the required resilient enablers among the AFSC stakeholders.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104225"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142841247","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

BlurRes-UNet: A novel neural network for automated surface characterisation in metrology BlurRes-UNet：一种新的神经网络，用于计量中的自动表面表征

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104228

Weixin Cui , Shan Lou , Wenhan Zeng , Visakan Kadirkamanathan , Yuchu Qin , Paul J. Scott , Xiangqian Jiang

Surface characterisation is essential in metrology for precise measurement and analysis of surface features, ensuring product quality and compliance with industry standards. Form removal is the primary step in surface characterisation, isolating features of interest by eliminating the primary shape from measurements. Traditional least-squares methods, as specified in ISO standards, are effective but offer limited adaptability for diverse surfaces and often require manual parameter tuning. With this limitation in mind, this paper proposes BlurRes-UNet, a deep learning-based model designed for fully automatic form removal. Built on an encoder–decoder architecture with residual learning, skip connections, and a tailored loss function, the model incorporates domain knowledge, feature engineering, and regularisation techniques to optimise performance with limited training data. The model is evaluated against traditional least squares methods and assessed using various strategies to demonstrate its performance and robustness. It processes surfaces of 256 × 256 resolution in 7.32 ms per sample on a T4 GPU, achieving superior accuracy in recognising reference forms across diverse surfaces compared to traditional methods. The results suggest that the model is capable of accurately recognising different order reference forms from diverse surfaces, facilitating an autonomous surface characterisation system without the need for manual intervention.

表面表征对于精确测量和分析表面特征，确保产品质量和符合行业标准的计量学至关重要。形状去除是表面表征的主要步骤，通过消除测量中的主要形状来隔离感兴趣的特征。ISO标准中规定的传统最小二乘方法是有效的，但对不同表面的适应性有限，并且通常需要手动调整参数。考虑到这一限制，本文提出了BlurRes-UNet，这是一种基于深度学习的模型，旨在实现全自动表单删除。该模型建立在具有残差学习、跳过连接和定制损失函数的编码器-解码器架构上，结合了领域知识、特征工程和正则化技术，以有限的训练数据优化性能。利用传统的最小二乘法对模型进行了评估，并使用各种策略对模型进行了评估，以证明其性能和鲁棒性。它在T4 GPU上以7.32 ms的速度处理256 × 256分辨率的表面，与传统方法相比，在识别不同表面的参考形式方面取得了卓越的准确性。结果表明，该模型能够准确地识别来自不同表面的不同顺序的参考形式，促进自主表面表征系统，而无需人工干预。

{"title":"BlurRes-UNet: A novel neural network for automated surface characterisation in metrology","authors":"Weixin Cui , Shan Lou , Wenhan Zeng , Visakan Kadirkamanathan , Yuchu Qin , Paul J. Scott , Xiangqian Jiang","doi":"10.1016/j.compind.2024.104228","DOIUrl":"10.1016/j.compind.2024.104228","url":null,"abstract":"<div><div>Surface characterisation is essential in metrology for precise measurement and analysis of surface features, ensuring product quality and compliance with industry standards. Form removal is the primary step in surface characterisation, isolating features of interest by eliminating the primary shape from measurements. Traditional least-squares methods, as specified in ISO standards, are effective but offer limited adaptability for diverse surfaces and often require manual parameter tuning. With this limitation in mind, this paper proposes BlurRes-UNet, a deep learning-based model designed for fully automatic form removal. Built on an encoder–decoder architecture with residual learning, skip connections, and a tailored loss function, the model incorporates domain knowledge, feature engineering, and regularisation techniques to optimise performance with limited training data. The model is evaluated against traditional least squares methods and assessed using various strategies to demonstrate its performance and robustness. It processes surfaces of 256 × 256 resolution in 7.32 ms per sample on a T4 GPU, achieving superior accuracy in recognising reference forms across diverse surfaces compared to traditional methods. The results suggest that the model is capable of accurately recognising different order reference forms from diverse surfaces, facilitating an autonomous surface characterisation system without the need for manual intervention.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104228"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142911787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A robotic skill transfer learning framework of dynamic manipulation for fabric placement 织物放置动态操作的机器人技能迁移学习框架

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104216

Tianyu Fu , Cheng Li , Yunfeng Bai , Fengming Li , Jiang Wu , Chaoqun Wang , Rui Song

Placing fabric poses a challenge to robots since fabric with high dimensional configuration space can deform during manipulation. Existing methods for placing fabric mostly rely on static operations, which are inefficient and require a large workspace. Therefore, this study applies dynamic manipulation (manipulating uncontrollable parts of the fabric by swinging) to fabric placement, proposing a novel learning framework for robotic dynamic fabric placement skill learning and generalization. The proposed framework integrates reinforcement learning with imitation learning, leveraging expert demonstration data to guide and accelerate skill acquisition. Additionally, fabric characteristics are combined with imitation learning to enable the transfer and generalization of the learned policy to real-world environments The experiments suggest that the proposed framework is capable of achieving the placement tasks for a range of positions and fabrics. For success rate, the policy of the proposed framework ultimately achieves a flatness of exceeding 95% and a placement distance error of less than 2 mm. Moreover, the proposed approach is similar in operation time to the fastest method, while it can reduce the space required for manipulating the fabric by over 15%. Compared with other placement policies, it is promising because of its high accuracy, flexibility, efficiency, as well as adaptability.

由于高维构形空间的织物在操作过程中会发生变形，对机器人放置织物提出了挑战。现有的放置织物的方法大多依赖于静态操作，这种方法效率低下，并且需要很大的工作空间。因此，本研究将动态操作（通过摆动操纵织物的不可控部分）应用于织物放置，为机器人动态放置织物技能的学习和泛化提出了一种新的学习框架。该框架将强化学习与模仿学习相结合，利用专家演示数据来指导和加速技能习得。此外，将织物特征与模仿学习相结合，使学习策略能够迁移和推广到现实环境中。实验表明，所提出的框架能够实现一系列位置和织物的放置任务。在成功率方面，所提出的框架策略最终实现了超过95%的平面度和小于2mm的放置距离误差。此外，该方法在操作时间上与最快的方法相似，同时可以将操作织物所需的空间减少15%以上。与其他定位策略相比，该策略具有精度高、灵活性强、效率高、适应性强等优点。

{"title":"A robotic skill transfer learning framework of dynamic manipulation for fabric placement","authors":"Tianyu Fu , Cheng Li , Yunfeng Bai , Fengming Li , Jiang Wu , Chaoqun Wang , Rui Song","doi":"10.1016/j.compind.2024.104216","DOIUrl":"10.1016/j.compind.2024.104216","url":null,"abstract":"<div><div>Placing fabric poses a challenge to robots since fabric with high dimensional configuration space can deform during manipulation. Existing methods for placing fabric mostly rely on static operations, which are inefficient and require a large workspace. Therefore, this study applies dynamic manipulation (manipulating uncontrollable parts of the fabric by swinging) to fabric placement, proposing a novel learning framework for robotic dynamic fabric placement skill learning and generalization. The proposed framework integrates reinforcement learning with imitation learning, leveraging expert demonstration data to guide and accelerate skill acquisition. Additionally, fabric characteristics are combined with imitation learning to enable the transfer and generalization of the learned policy to real-world environments The experiments suggest that the proposed framework is capable of achieving the placement tasks for a range of positions and fabrics. For success rate, the policy of the proposed framework ultimately achieves a flatness of exceeding 95% and a placement distance error of less than 2 mm. Moreover, the proposed approach is similar in operation time to the fastest method, while it can reduce the space required for manipulating the fabric by over 15%. Compared with other placement policies, it is promising because of its high accuracy, flexibility, efficiency, as well as adaptability.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104216"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142763009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A context-aware decision support system for selecting explainable artificial intelligence methods in business organizations 在商业组织中选择可解释的人工智能方法的上下文感知决策支持系统

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104233

Marcelo I. Reis , João N.C. Gonçalves , Paulo Cortez , M. Sameiro Carvalho , João M. Fernandes

Explainable Artificial Intelligence (XAI) methods are valuable tools for promoting understanding, trust, and efficient use of Artificial Intelligence (AI) systems in business organizations. However, the question of how organizations should select suitable XAI methods for a given task and business context remains a challenge, particularly when the number of methods available in the literature continues to increase. Here, we propose a context-aware decision support system (DSS) to select, from a given set of XAI methods, those with higher suitability to the needs of stakeholders operating in a given AI-based business problem. By including the human-in-the-loop, our DSS comprises an application-grounded analytical metric designed to facilitate the selection of XAI methods that align with the business stakeholders’ desiderata and promote a deeper understanding of the results generated by a given machine learning model. The proposed system was tested on a real supply chain demand problem, using real data and real users. The results provide evidence on the usefulness of our metric in selecting XAI methods based on the feedback and analytical maturity of stakeholders from the deployment context. We believe that our DSS is sufficiently flexible and understandable to be applied in a variety of business contexts, with stakeholders with varying degrees of AI literacy.

可解释的人工智能（XAI）方法是促进商业组织中人工智能（AI）系统的理解、信任和有效使用的宝贵工具。然而，组织应该如何为给定的任务和业务上下文选择合适的XAI方法的问题仍然是一个挑战，特别是当文献中可用的方法数量继续增加时。在这里，我们提出了一个上下文感知的决策支持系统（DSS），从给定的一组XAI方法中选择那些更适合在给定的基于人工智能的业务问题中操作的利益相关者的需求的方法。通过包括人在循环，我们的DSS包括一个基于应用程序的分析度量，旨在促进XAI方法的选择，使其与业务利益相关者的期望保持一致，并促进对给定机器学习模型生成的结果的更深入理解。该系统在一个真实的供应链需求问题上进行了测试，使用了真实数据和真实用户。结果证明了我们的度量在基于来自部署上下文的涉众的反馈和分析成熟度选择XAI方法时的有用性。我们相信，我们的决策支持系统具有足够的灵活性和可理解性，可以应用于各种商业环境，以及具有不同程度人工智能素养的利益相关者。

{"title":"A context-aware decision support system for selecting explainable artificial intelligence methods in business organizations","authors":"Marcelo I. Reis , João N.C. Gonçalves , Paulo Cortez , M. Sameiro Carvalho , João M. Fernandes","doi":"10.1016/j.compind.2024.104233","DOIUrl":"10.1016/j.compind.2024.104233","url":null,"abstract":"<div><div>Explainable Artificial Intelligence (XAI) methods are valuable tools for promoting understanding, trust, and efficient use of Artificial Intelligence (AI) systems in business organizations. However, the question of how organizations should select suitable XAI methods for a given task and business context remains a challenge, particularly when the number of methods available in the literature continues to increase. Here, we propose a context-aware decision support system (DSS) to select, from a given set of XAI methods, those with higher suitability to the needs of stakeholders operating in a given AI-based business problem. By including the human-in-the-loop, our DSS comprises an application-grounded analytical metric designed to facilitate the selection of XAI methods that align with the business stakeholders’ desiderata and promote a deeper understanding of the results generated by a given machine learning model. The proposed system was tested on a real supply chain demand problem, using real data and real users. The results provide evidence on the usefulness of our metric in selecting XAI methods based on the feedback and analytical maturity of stakeholders from the deployment context. We believe that our DSS is sufficiently flexible and understandable to be applied in a variety of business contexts, with stakeholders with varying degrees of AI literacy.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104233"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142911788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

YOLOv10-pose and YOLOv9-pose: Real-time strawberry stalk pose detection models YOLOv10-pose和YOLOv9-pose：实时草莓茎秆姿态检测模型

IF 8.2 1区计算机科学 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers in Industry

Pub Date : 2025-02-01 DOI: 10.1016/j.compind.2024.104231

Zhichao Meng , Xiaoqiang Du , Ranjan Sapkota , Zenghong Ma , Hongchao Cheng

In the computer-aided industry, particularly within the domain of agricultural automation, fruit pose detection is critical for optimizing efficiency across various applications such as robotic harvesting, aerial crop surveillance, precision pruning, and automated sorting. These technologies enhance productivity and precision, addressing challenges posed by an aging labor force and the increasing demand for sophisticated robotic applications in agriculture. This is particularly crucial for strawberries, which are globally recognized for their high nutritional value. The strawberry pickting robots generally cut the stems, so knowing the pose of the strawberry stalks before cutting can effectively adjust the pose of the end effector, thereby improving the success rate of picking. This paper referred to the keypoint detection branch and loss function of the YOLOv8-pose model, and combined the latest YOLOv9 and YOLOv10 object detection models to propose YOLOv9-pose and YOLOv10-pose. The experimental results showed that YOLOv9-base-pose had the best comprehensive performance, reaching 0.962 in Box_mAP50 and 0.914 in Pose_mAP50, and the speed met the real-time requirement of FPS 51. The entire YOLOv10-pose series did not achieve satisfactory accuracy, but not using non-maximum suppression did indeed speed up the post-processing. In the YOLOv10-pose series, YOLOv10m-pose achieved the best comprehensive performance with Box_mAP50 of 0.954, Pose_ mAP50 of 0.903, and a speed of 61 FPS. Comparing YOLOv9-base-pose with the entire series of YOLOv8-pose and YOLOv5-pose also demonstrated the superior performance of YOLOv9-base-pose. YOLOv9-pose and YOLOv10-pose can provide a theoretical basis for pose detection and a reference for other similar fruit pose detection.

在计算机辅助工业中，特别是在农业自动化领域，水果姿态检测对于优化各种应用的效率至关重要，例如机器人收获、空中作物监视、精确修剪和自动分拣。这些技术提高了生产率和精度，解决了劳动力老龄化和农业对复杂机器人应用日益增长的需求所带来的挑战。这对全球公认的高营养价值草莓来说尤其重要。草莓采摘机器人一般会对草莓茎进行切割，因此在切割前了解草莓茎的姿态可以有效地调整末端执行器的姿态，从而提高采摘的成功率。本文参考了YOLOv8-pose模型的关键点检测分支和损失函数，结合最新的YOLOv9和YOLOv10目标检测模型，提出了YOLOv9-pose和YOLOv10-pose。实验结果表明，YOLOv9-base-pose综合性能最好，在Box_mAP50中达到0.962，在Pose_mAP50中达到0.914，速度满足FPS 51的实时性要求。整个YOLOv10-pose系列没有达到令人满意的精度，但不使用非最大抑制确实加快了后处理速度。在YOLOv10-pose系列中，YOLOv10m-pose的综合性能最好，Box_mAP50为0.954，Pose_ mAP50为0.903，速度为61 FPS。将YOLOv9-base-pose与YOLOv8-pose和YOLOv5-pose的整个系列进行比较，也证明了YOLOv9-base-pose的优越性能。YOLOv9-pose和YOLOv10-pose可以为姿态检测提供理论基础，也可以为其他类似水果的姿态检测提供参考。

{"title":"YOLOv10-pose and YOLOv9-pose: Real-time strawberry stalk pose detection models","authors":"Zhichao Meng , Xiaoqiang Du , Ranjan Sapkota , Zenghong Ma , Hongchao Cheng","doi":"10.1016/j.compind.2024.104231","DOIUrl":"10.1016/j.compind.2024.104231","url":null,"abstract":"<div><div>In the computer-aided industry, particularly within the domain of agricultural automation, fruit pose detection is critical for optimizing efficiency across various applications such as robotic harvesting, aerial crop surveillance, precision pruning, and automated sorting. These technologies enhance productivity and precision, addressing challenges posed by an aging labor force and the increasing demand for sophisticated robotic applications in agriculture. This is particularly crucial for strawberries, which are globally recognized for their high nutritional value. The strawberry pickting robots generally cut the stems, so knowing the pose of the strawberry stalks before cutting can effectively adjust the pose of the end effector, thereby improving the success rate of picking. This paper referred to the keypoint detection branch and loss function of the YOLOv8-pose model, and combined the latest YOLOv9 and YOLOv10 object detection models to propose YOLOv9-pose and YOLOv10-pose. The experimental results showed that YOLOv9-base-pose had the best comprehensive performance, reaching 0.962 in Box_mAP50 and 0.914 in Pose_mAP50, and the speed met the real-time requirement of FPS 51. The entire YOLOv10-pose series did not achieve satisfactory accuracy, but not using non-maximum suppression did indeed speed up the post-processing. In the YOLOv10-pose series, YOLOv10m-pose achieved the best comprehensive performance with Box_mAP50 of 0.954, Pose_ mAP50 of 0.903, and a speed of 61 FPS. Comparing YOLOv9-base-pose with the entire series of YOLOv8-pose and YOLOv5-pose also demonstrated the superior performance of YOLOv9-base-pose. YOLOv9-pose and YOLOv10-pose can provide a theoretical basis for pose detection and a reference for other similar fruit pose detection.</div></div>","PeriodicalId":55219,"journal":{"name":"Computers in Industry","volume":"165 ","pages":"Article 104231"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142884318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0