Article
Chemistry, Analytical
Paolo Brambilla, Chiara Conese, Davide Maria Fabris, Paolo Chiariotti, Marco Tarabini
Summary: Quality inspection in industrial production is benefiting from the combination of vision-based techniques and artificial intelligence algorithms. This paper discusses the defect identification problem for circularly symmetric mechanical components and compares the performances of a standard algorithm with a Deep Learning (DL) approach. The standard algorithm provides better results in terms of accuracy and computational time, but DL achieves high accuracy in identifying damaged teeth. The possibility of extending the methods and results to other circularly symmetrical components is also analyzed and discussed.
Article
Computer Science, Artificial Intelligence
Yuda Song, Zhuqing He, Hui Qian, Xin Du
Summary: Image dehazing is a low-level vision task to estimate haze-free images. Convolutional neural network methods dominate this task, but vision Transformers haven't made a breakthrough. This study introduces DehazeFormer, an improved version of Swin Transformer, with modified normalization layer, activation function, and spatial information aggregation. Multiple variants of DehazeFormer were trained and shown to be effective.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Construction & Building Technology
E. Aldao, L. Fernandez-Pardo, L. M. Gonzalez-deSantos, H. Gonzalez-Jorge
Summary: This work proposes and compares different methods for autonomously inspecting railway bolts and clips. A prototype autonomous data acquisition system was developed, using LiDAR and camera sensors to obtain information about the state of the railway track. The system was tested on a railway track at the University of Vigo, and images were processed using image segmentation algorithms and a neural network to detect the bolts and clips.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Computer Science, Artificial Intelligence
Nimel Sworna Ross, Paul T. T. Sheeba, C. Sherin Shibi, Munish Kumar Gupta, Mehmet Erdi Korkmaz, Vishal S. Sharma
Summary: The condition of cutting tools is crucial in metal cutting. Predictive models based on machine learning have been proposed to anticipate and avoid tool failures. This study introduces the use of transfer learning technology to detect tool wear under different cutting environments. Pre-trained networks are used to determine the tool condition, and the best-performing network is recommended for tool condition monitoring, considering hyperparameters. This methodology can be highly helpful for classifying and suggesting suitable cutting conditions, especially with limited data.
JOURNAL OF INTELLIGENT MANUFACTURING
(2023)
Article
Automation & Control Systems
Linzhi Xia, Yizhu Shi, Hongjie Lin, Houyuan Zheng, Xincheng Cao, Binqiang Chen, Yuqing Zhou, Weifang Sun
Summary: This paper proposes an improved vision-based method for segmentation and quantitative evaluation of tool wear monitoring. Experimental results show its effectiveness.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2022)
Article
Computer Science, Interdisciplinary Applications
Hao Guo, Yu Zhang, Kunpeng Zhu
Summary: This study proposes a multi-scale pyramid attention network (MPAN) for tool wear monitoring. The method can accurately monitor tool wear and provide interpretability from the aspects of network structure design and feature extraction. Experimental results demonstrate the effectiveness and feasibility of this method.
COMPUTERS IN INDUSTRY
(2022)
Article
Chemistry, Multidisciplinary
Aleksandar Miltenovic, Ivan Rakonjac, Alexandru Oarcea, Marko Peric, Damjan Rangelov
Summary: In this paper, a machine vision system is proposed for detecting and monitoring pitting as part of the inspection process, ensuring the proper operation of gears. The implemented system uses a faster R-CNN network for detection and positioning, achieving high accuracy in pitting damage detection.
APPLIED SCIENCES-BASEL
(2022)
Article
Engineering, Multidisciplinary
Yufeng Shu, Bin Li, Hui Lin
Summary: The surface quality safety inspection of LED chips is essential in production, with traditional methods struggling to keep up with shrinking chip sizes. Deep convolutional neural networks have made significant breakthroughs in this field, surpassing traditional models in performance.
Article
Chemistry, Multidisciplinary
Hossam A. Gabbar, Abderrazak Chahid, Md. Jamiul Alam Khan, Oluwabukola Grace Adegboro, Matthew Immanuel Samson
Summary: In this paper, a new toolbox called CT-Based Integrity Monitoring System (CTIMS-Toolbox) is proposed for automated inspection of CT images and volumes in non-destructive testing for industrial tool quality and safety control. The toolbox consists of three main modules: database management, pre-processing, and defect inspection, utilizing computer vision and deep learning techniques.
APPLIED SCIENCES-BASEL
(2022)
Article
Engineering, Multidisciplinary
Xianyi Zhai, Meng Huang, Honglei Wei
Summary: To address the problem of chip location recognition, this paper proposes a lightweight E-YOLOv5 based chip detection algorithm, which is based on the You Only Look Once version 5 (YOLOv5s) algorithm. The algorithm enhances the model robustness by using a simulated exposure algorithm and reduces the model size by introducing EfficientNet. Furthermore, the Selective Kernel Neural Network (SKNet) module is introduced to enhance the feature extraction ability and improve the training efficiency. The experimental results demonstrate that the proposed algorithm outperforms other algorithms in terms of precision, recall rate, loss value, model size, training time, and image processing speed.
ENGINEERING RESEARCH EXPRESS
(2023)
Review
Computer Science, Artificial Intelligence
Alzayat Saleh, Marcus Sheaves, Dean Jerry, Mostafa Rahimi Azghadi
Summary: This paper introduces the application of deep learning in underwater fish habitat monitoring. The tutorial explains the key concepts of DL and provides a step-by-step procedure for developing DL algorithms. Additionally, the paper surveys underwater fish datasets and various DL techniques, and discusses the challenges and opportunities in the field.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Engineering, Civil
Hamed Haghighi, Mehrdad Dianati, Valentina Donzella, Kurt Debattista
Summary: Camera image simulation is crucial for virtual validation of autonomous vehicles and robots, as well as for creating image datasets for training vision models. To address the computational complexity, we propose a technique based on Stereo Super Resolution (SSR) to speed up the simulation of stereo images, achieving promising results in terms of speed and performance.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Chemistry, Multidisciplinary
Mateusz Dziubek, Jacek Rysinski, Daniel Jancarczyk
Summary: This study explores the application of machine vision and ViDiDetect in assessing cutting tool wear. Machine vision systems offer a non-contact and non-destructive approach to evaluation by capturing high-resolution images and analyzing wear patterns. The investigation demonstrates the potential of machine vision and ViDiDetect in automating cutting tool wear assessment.
APPLIED SCIENCES-BASEL
(2023)
Article
Chemistry, Analytical
Yaping Zhang, Xiaozhi Qi, Tao Wang, Yuanhang He
Summary: Tool wear condition monitoring is crucial for mechanical processing automation, and this paper proposes a new deep learning model to accurately identify tool wear status. By transforming the force signal into a two-dimensional image using CWT, STFT, and GASF methods, the images are then analyzed using a CNN model. The results show that this method achieves an accuracy above 90%, outperforming other models like AlexNet and ResNet. The images generated by the CWT method exhibit the highest accuracy in tool wear state recognition due to its ability to extract local features and resist noise. These findings highlight the potential advantages of using a force signal transformed into a two-dimensional image and applying CNN models in tool wear state recognition, indicating promising applications in industrial production.
Article
Engineering, Industrial
Boling Yan, Lida Zhu, Yichao Dun
Summary: The study presents a real-time tool wear monitoring and prediction technique using deep learning, which is more efficient and accurate compared to traditional feature extraction methods, with a maximum prediction error of around 8 micrometers.
JOURNAL OF MANUFACTURING SYSTEMS
(2021)
Article
Automation & Control Systems
Nicola Milan, Paolo Parenti, Massimiliano Annoni, Marco Sorgato, Giovanni Lucchetta
Summary: An innovative process chain based on carbide tools micromilling of mould gratings was developed for mass production of diffractive patterns on injection moulded parts. A micromilling experimental campaign was conducted on a nickel-phosphorus (NiP) thick coating to evaluate the influence of cutting parameters on diffractive surface quality, followed by replication on ABS, PC, and PMMA by injection moulding. Results indicate that the proposed process chain is suitable for low-cost mass production of polymeric parts with diffractive microstructures.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2021)
Article
Engineering, Mechanical
Shubhavardhan Ramadurga Narasimharaju, Weidong Liu, Wenhan Zeng, Tian Long See, Paul Scott, Xiangqian Jane Jiang, Shan Lou
Summary: The study systematically investigates the impact of varying surface inclination angles on the surface quality of additively manufactured parts and finds that the surface topography is strongly correlated with the surface inclination angles.
JOURNAL OF TRIBOLOGY-TRANSACTIONS OF THE ASME
(2021)
Article
Engineering, Mechanical
S. Lou, Z. Zhu, W. Zeng, C. Majewski, P. J. Scott, X. Jiang
Summary: This study investigates the material ratio curves of surfaces produced by additive manufacturing processes and finds that these curves vary depending on the specific process and surface orientations. The study also highlights the effectiveness of using the material ratio curve as an analysis tool to differentiate various AM surface topographies.
SURFACE TOPOGRAPHY-METROLOGY AND PROPERTIES
(2021)
Article
Automation & Control Systems
Peizhi Shi, Qunfen Qi, Yuchu Qin, Paul J. Scott, Xiangqian Jiang
Summary: In this study, a novel deep learning approach named SsdNet is proposed to tackle the machining feature localization and recognition problem, achieving state-of-the-art performance in feature recognition and localization. The method modifies the network architecture and output of SSD, and utilizes advanced techniques to enhance recognition performance.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Engineering, Multidisciplinary
Stefano Petro, Luca Pagani, Giovanni Moroni, Paul J. Scott
Summary: Additive Manufacturing (AM) revolutionizes manufacturing by enabling complex geometries; traditional GPS/GD&T practices lack flexibility in specifying and verifying geometric tolerances; volumetric representations and X-ray computed tomography (XCT) are essential for accurately measuring complex parts.
PRECISION ENGINEERING-JOURNAL OF THE INTERNATIONAL SOCIETIES FOR PRECISION ENGINEERING AND NANOTECHNOLOGY
(2021)
Article
Computer Science, Interdisciplinary Applications
Peizhi Shi, Qunfen Qi, Yuchu Qin, Paul J. Scott, Xiangqian Jiang
Summary: In the field of intelligent manufacturing, recognizing interacting features on a CAD model is a critical yet challenging task. Some learning methods struggle with highly interacting features and require a large number of 3D models for training. The proposed method RDetNet can effectively recognize highly interacting features with small training samples.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Engineering, Manufacturing
Paolo Parenti, Dario Puccio, Bianca Maria Colosimo, Quirico Semeraro
Summary: This paper presents a novel solution for assessing the printability of complex geometries in metal additive manufacturing. The approach combines logistic regression and quality data to evaluate the likelihood of producing defect-free complex geometries. The paper also investigates the printing capability of a new emerging metal additive manufacturing technology.
JOURNAL OF MANUFACTURING PROCESSES
(2022)
Article
Multidisciplinary Sciences
Q. Qi, W. Terkaj, M. Urgo, X. Jiang, P. J. Scott
Summary: As manufacturing undergoes the Industry 4.0 revolution, the complexity of digitized manufacturing systems increases due to a large amount of data and information exchange. To address this complexity, one solution is to design and operate digital twin models at different levels of abstraction and detail. To enable efficient information flow between models with different levels of detail, a mathematical structure known as a delta lens has been developed. A hybrid delta lens has also been proposed to support different types of abstractions in manufacturing.
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2022)
Article
Automation & Control Systems
Edoardo Copertaro, Massimiliano Annoni
Summary: This paper investigates the correlation between the airborne acoustic emission of an abrasive waterjet cutting head and the jet kinetic power. By conducting factorial studies, different kinetic powers are obtained by firing the jet at various water pressures and abrasive feed rates. The acoustic emission data at frequencies above 40 kHz is found to be a robust indicator of the jet kinetic power and its pressure-induced variations.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2022)
Article
Chemistry, Analytical
Margherita Pizzi, Francesco De Gaetano, Marco Ferroni, Federica Boschetti, Massimiliano Annoni
Summary: The mechanisms of microdrilling of pure Mg material for intraocular drug delivery device prototyping were experimentally studied. The experiments were conducted with different microdrills and cutting parameters, and the results showed that the burr height is not uniform along the circumference of the holes and the roughness of the inner surface increases with higher feed rate.
Article
Automation & Control Systems
Yuchu Qin, Qunfen Qi, Peizhi Shi, Paul J. J. Scott, Xiangqian Jiang
Summary: In this paper, an approach for material selection in metal additive manufacturing based on three-way decision-making is proposed. The approach is divided into three stages, including establishing a decision matrix, calculating summary loss function and conditional probability, and generating three-way decision results. The effectiveness of the approach is demonstrated through a material selection example and a comparison with existing approaches. The results suggest that the proposed approach is as effective as existing approaches and more flexible and advantageous in solving material selection problems in metal additive manufacturing.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Yuchu Qin, Qunfen Qi, Peizhi Shi, Paul J. Scott, Xiangqian Jiang
Summary: In this paper, a weighted averaging operator of linguistic interval-valued intuitionistic fuzzy numbers (LIVIFNs) based on Dempster-Shafer evidence theory is proposed for solving cognitively inspired decision-making problems. The developed operational rules of LIVIFNs are proven to be always invariant and persistent, and the constructed aggregation operator is proven to be always monotone. The effectiveness and advantage of the presented method are demonstrated through quantitative comparisons with several existing methods.
COGNITIVE COMPUTATION
(2023)
Article
Engineering, Manufacturing
Talha Sunar, Paolo Parenti, Tansel Tuncay, Dursun Ozyurek, Massimiliano Annoni
Summary: Improving scientific knowledge around the manufacturing of nanocomposites is crucial due to their wide range of applications. Powder metallurgy is a reliable process for producing these materials, but machining postprocessing is often necessary for achieving specific tolerances and quality requirements.
JOURNAL OF MANUFACTURING AND MATERIALS PROCESSING
(2023)
Article
Engineering, Industrial
Xiangqian Jiang, Nicola Senin, Paul J. Scott, Francois Blateyron
Summary: This keynote paper provides an overview of emerging technologies for feature-based characterisation of surface topographies. It introduces the concept of the feature spectrum and discusses its applications, guidelines for future industrial use, and considerations for future challenges.
CIRP ANNALS-MANUFACTURING TECHNOLOGY
(2021)
Article
Engineering, Industrial
Walter Terkaj, Qunfen Qi, Marcello Urgo, Paul J. Scott, Xiangqian Jiang
Summary: By leveraging ontologies and delta-lenses, this study enables multiscale models of manufacturing systems to map digital models and flow data while assessing the capability of lower-detail models to approximate system behavior, ultimately deciding the positions of sensors on an assembly line.
CIRP ANNALS-MANUFACTURING TECHNOLOGY
(2021)