Sensors

25 pages, 4786 KiB

Open AccessArticle

Air Pollution Measurement and Dispersion Simulation Using Remote and In Situ Monitoring Technologies in an Industrial Complex in Busan, South Korea

by Naghmeh Dehkhoda, Juhyeon Sim, Juseon Shin, Sohee Joo, Sung Hwan Cho, Jeong Hun Kim and Youngmin Noh

Sensors 2024, 24(23), 7836; https://doi.org/10.3390/s24237836 (registering DOI) - 7 Dec 2024

Abstract

Rapid industrialization and the influx of human resources have led to the establishment of industrial complexes near urban areas, exposing residents to various air pollutants. This has led to a decline in air quality, impacting neighboring residential areas adversely, which highlights the urgent [...] Read more.

Rapid industrialization and the influx of human resources have led to the establishment of industrial complexes near urban areas, exposing residents to various air pollutants. This has led to a decline in air quality, impacting neighboring residential areas adversely, which highlights the urgent need to monitor air pollution in these areas. Recent advancements in technology, such as Solar Occultation Flux (SOF) and Sky Differential Optical Absorption Spectroscopy (SkyDOAS) used as remote sensing techniques and mobile extraction Fourier Transform Infrared Spectrometry (MeFTIR) used as an in situ technique, now offer enhanced precision in estimating the pollutant emission flux and identifying primary sources. In a comprehensive study conducted in 2020 in the Sinpyeong Jangrim Industrial Complex in Busan City, South Korea, a mobile laboratory equipped with SOF, SkyDOAS, and MeFTIR technologies was employed to approximate the emission flux of total alkanes, sulfur dioxide (SO₂), nitrogen dioxide (NO₂), formaldehyde (HCHO), and methane (CH₄). Using the HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) diffusion model, pollutant dispersion to residential areas was simulated. The highest average daily emission flux was observed for total alkanes, with values of 69.9 ± 71.6 kg/h and 84.1 ± 85.8 kg/h in zones S₁ and S₂ of the Sinpyeong Jangrim Industrial Complex, respectively. This is primarily due to the prevalence of metal manufacturing and mechanical equipment industries in the area. The HYSPLIT diffusion model confirmed elevated pollution levels in residential areas located southeast of the industrial complex, underscoring the influence of the dominant northwesterly wind direction and wind speed on pollutant dispersion. This highlights the urgent need for targeted interventions to address and mitigate air pollution in downwind residential areas. The total annual emission fluxes were estimated at 399,984 kg/yr and 398,944 kg/yr for zones S₁ and S₂, respectively. A comparison with the Pollutant Release and Transfer Registers (PRTRs) survey system revealed that the total annual emission fluxes in this study were approximately 24.3 and 4.9 times higher than those reported by PRTRs. This indicates a significant underestimation of the impact of small businesses on local air quality, which was not accounted for in the PRTR survey system. Full article

(This article belongs to the Special Issue Remote Sensing in Atmospheric Measurements)

► Show Figures

Figure 1

19 pages, 1953 KiB

Open AccessArticle

An Efficient Acoustic Camera Simulator for Robotic Applications: Design, Implementation, and Validation

by Jisung Park

Sensors 2024, 24(23), 7835; https://doi.org/10.3390/s24237835 (registering DOI) - 7 Dec 2024

Abstract

As forward-looking imaging sonar, also called an acoustic camera, has emerged as an important sensor for marine robotics applications, its simulators have attracted considerable research attention within the field. This paper presents an acoustic camera simulator that efficiently generates acoustic images using only [...] Read more.

As forward-looking imaging sonar, also called an acoustic camera, has emerged as an important sensor for marine robotics applications, its simulators have attracted considerable research attention within the field. This paper presents an acoustic camera simulator that efficiently generates acoustic images using only the depth information of the scene. The simulator approximates the acoustic beam of a real acoustic camera as a set of acoustic rays originating from the center of the acoustic camera. A simplified active sonar model and error models for the acoustic rays are incorporated to account for the environmental factors which can affect the rays. The simulator is implemented in the Gazebo simulator, a robotic simulator, and does not rely on any other components except depth rendering. To evaluate the performance of the developed simulator, qualitative and quantitative comparisons were conducted between simulated images and actual acoustic images obtained in a test tank. Full article

(This article belongs to the Section Sensors and Robotics)

► Show Figures

Figure 1

28 pages, 2265 KiB

Open AccessArticle

TUH-NAS: A Triple-Unit NAS Network for Hyperspectral Image Classification

by Feng Chen, Baishun Su and Zongpu Jia

Sensors 2024, 24(23), 7834; https://doi.org/10.3390/s24237834 (registering DOI) - 7 Dec 2024

Abstract

Over the last few years, neural architecture search (NAS) technology has achieved good results in hyperspectral image classification. Nevertheless, existing NAS-based classification methods have not specifically focused on the complex connection between spectral and spatial data. Strengthening the integration of spatial and spectral [...] Read more.

Over the last few years, neural architecture search (NAS) technology has achieved good results in hyperspectral image classification. Nevertheless, existing NAS-based classification methods have not specifically focused on the complex connection between spectral and spatial data. Strengthening the integration of spatial and spectral features is crucial to boosting the overall classification efficacy of hyperspectral images. In this paper, a triple-unit hyperspectral NAS network (TUH-NAS) aimed at hyperspectral image classification is introduced, where the fusion unit emphasizes the enhancement of the intrinsic relationship between spatial and spectral information. We designed a new hyperspectral image attention mechanism module to increase the focus on critical regions and enhance sensitivity to priority areas. We also adopted a composite loss function to enhance the model's focus on hard-to-classify samples. Experimental evaluations on three publicly accessible hyperspectral datasets demonstrated that, despite utilizing a limited number of samples, TUH-NAS outperforms existing NAS classification methods in recognizing object boundaries. Full article

(This article belongs to the Section Sensing and Imaging)

14 pages, 1158 KiB

Open AccessArticle

Extreme R-CNN: Few-Shot Object Detection via Sample Synthesis and Knowledge Distillation

by Shenyong Zhang, Wenmin Wang, Zhibing Wang, Honglei Li, Ruochen Li and Shixiong Zhang

Sensors 2024, 24(23), 7833; https://doi.org/10.3390/s24237833 (registering DOI) - 7 Dec 2024

Abstract

Traditional object detectors require extensive instance-level annotations for training. Conversely, few-shot object detectors, which are generally fine-tuned using limited data from unknown classes, tend to show biases toward base categories and are susceptible to variations within these unknown samples. To mitigate these challenges, [...] Read more.

Traditional object detectors require extensive instance-level annotations for training. Conversely, few-shot object detectors, which are generally fine-tuned using limited data from unknown classes, tend to show biases toward base categories and are susceptible to variations within these unknown samples. To mitigate these challenges, we introduce a Two-Stage Fine-Tuning Approach (TFA) named Extreme R-CNN, designed to operate effectively with extremely limited original samples through the integration of sample synthesis and knowledge distillation. Our approach involves synthesizing new training examples via instance clipping and employing various data-augmentation techniques. We enhance the Faster R-CNN architecture by decoupling the regression and classification components of the Region of Interest (RoI), allowing synthetic samples to train the classification head independently of the object-localization process. Comprehensive evaluations on the Microsoft COCO and PASCAL VOC datasets demonstrate significant improvements over baseline methods. Specifically, on the PASCAL VOC dataset, the average precision for novel categories is enhanced by up to 15 percent, while on the more complex Microsoft COCO benchmark it is enhanced by up to 6.1 percent. Remarkably, in the 1-shot scenario, the AP50 of our model exceeds that of the baseline model in the 10-shot setting within the PASCAL VOC dataset, confirming the efficacy of our proposed method. Full article

(This article belongs to the Collection Artificial Intelligence in Sensors Technology)

► Show Figures

Figure 1

16 pages, 4785 KiB

Open AccessArticle

Room-Temperature Ammonia Sensing Using Polyaniline-Coated Laser-Induced Graphene

by José Carlos Santos-Ceballos, Foad Salehnia, Frank Güell, Alfonso Romero, Xavier Vilanova and Eduard Llobet

Sensors 2024, 24(23), 7832; https://doi.org/10.3390/s24237832 (registering DOI) - 7 Dec 2024

Abstract

The reliable detection of ammonia at room temperature is crucial for not only maintaining environmental safety but also for reducing the risks of hazardous pollutants. In this study, the electrochemical modification of laser-induced graphene (LIG) with polyaniline (PANI) led to the development of [...] Read more.

The reliable detection of ammonia at room temperature is crucial for not only maintaining environmental safety but also for reducing the risks of hazardous pollutants. In this study, the electrochemical modification of laser-induced graphene (LIG) with polyaniline (PANI) led to the development of a chemo-resistive nanocomposite (PANI@LIG) for detecting ammonia levels at room temperature. The composite is characterized by field emission scanning electron microscopy, Fourier transforms infrared, and Raman and X-ray photoelectron spectroscopy. This work marks the first utilization of PANI@LIG for gas sensing and introduces a simple but effective approach for fabricating low-cost wearable gas sensors with high sensitivity and flexibility. Full article

(This article belongs to the Special Issue Recent Advances in Sensors for Chemical Detection Applications)

► Show Figures

Figure 1

18 pages, 1732 KiB

Open AccessArticle

A One-Dimensional Depthwise Separable Convolutional Neural Network for Bearing Fault Diagnosis Implemented on FPGA

by Yu-Pei Liang, Hao Chen and Ching-Che Chung

Sensors 2024, 24(23), 7831; https://doi.org/10.3390/s24237831 (registering DOI) - 7 Dec 2024

Abstract

This paper presents a hardware implementation of a one-dimensional convolutional neural network using depthwise separable convolution (DSC) on the VC707 FPGA development board. The design processes the one-dimensional rolling bearing current signal dataset provided by Paderborn University (PU), employing minimal preprocessing to maximize [...] Read more.

This paper presents a hardware implementation of a one-dimensional convolutional neural network using depthwise separable convolution (DSC) on the VC707 FPGA development board. The design processes the one-dimensional rolling bearing current signal dataset provided by Paderborn University (PU), employing minimal preprocessing to maximize the comprehensiveness of feature extraction. To address the high parameter demands commonly associated with convolutional neural networks (CNNs), the model incorporates DSC, significantly reducing computational complexity and parameter load. Additionally, the DoReFa-Net quantization method is applied to compress network parameters and activation function outputs, thereby minimizing memory usage. The quantized DSC model requires approximately 22 KB of storage and performs 1,203,128 floating-point operations in total. The implementation achieves a power consumption of 527 mW at a clock frequency of 50 MHz, while delivering a fault diagnosis accuracy of 96.12%. Full article

(This article belongs to the Special Issue Feature Papers in Physical Sensors 2024)

► Show Figures

Figure 1

14 pages, 3315 KiB

Open AccessArticle

Using a Bodily Weight-Fat Scale for Cuffless Blood Pressure Measurement Based on the Edge Computing System

by Shing-Hong Liu, Bo-Yan Wu, Xin Zhu and Chiun-Li Chin

Sensors 2024, 24(23), 7830; https://doi.org/10.3390/s24237830 (registering DOI) - 7 Dec 2024

Abstract

Blood pressure (BP) measurement is a major physiological information for people with cardiovascular diseases, such as hypertension, heart failure, and atherosclerosis. Moreover, elders and patients with kidney disease and diabetes mellitus also are suggested to measure their BP every day. The cuffless BP [...] Read more.

Blood pressure (BP) measurement is a major physiological information for people with cardiovascular diseases, such as hypertension, heart failure, and atherosclerosis. Moreover, elders and patients with kidney disease and diabetes mellitus also are suggested to measure their BP every day. The cuffless BP measurement has been developed in the past 10 years, which is comfortable to users. Now, ballistocardiogram (BCG) and impedance plethysmogram (IPG) could be used to perform the cuffless BP measurement. Thus, the aim of this study is to realize edge computing for the BP measurement in real time, which includes measurements of BCG and IPG signals, digital signal process, feature extraction, and BP estimation by machine learning algorithm. This system measured BCG and IPG signals from a bodily weight-fat scale with the self-made circuits. The signals were filtered to reduce the noise and segmented by 2 s. Then, we proposed a flowchart to extract the parameter, pulse transit time (PTT), within each segment. The feature included two calibration-based parameters and one calibration-free parameter was used to estimate BP with XGBoost. In order to realize the system in STM32F756ZG NUCLEO development board, we limited the hyperparameters of XGBoost model, including maximum depth (max_depth) and tree number (n_estimators). Results show that the error of systolic blood pressure (SBP) and diastolic blood pressure (DBP) in server-based computing are 2.64 ± 9.71 mmHg and 1.52 ± 6.32 mmHg, and in edge computing are 2.2 ± 10.9 mmHg and 1.87 ± 6.79 mmHg. This proposed method significantly enhances the feasibility of bodily weight-fat scale in the BP measurement for effective utilization in mobile health applications. Full article

(This article belongs to the Special Issue Integration of Edge/Fog Artificial Intelligence into Smart Distributed Systems)

► Show Figures

Figure 1

19 pages, 18714 KiB

Open AccessArticle

Hardware Implementation for Triaxial Contact-Force Estimation from Stress Tactile Sensor Arrays: An Efficient Design Approach

by María-Luisa Pinto-Salamanca, Wilson-Javier Pérez-Holguín and José A. Hidalgo-López

Sensors 2024, 24(23), 7829; https://doi.org/10.3390/s24237829 (registering DOI) - 7 Dec 2024

Abstract

This paper presents a contribution to the state of the art in the design of tactile sensing algorithms that take advantage of the characteristics of generalized sparse matrix-vector multiplication to reduce the area, power consumption, and data storage required for real-time hardware implementation. [...] Read more.

This paper presents a contribution to the state of the art in the design of tactile sensing algorithms that take advantage of the characteristics of generalized sparse matrix-vector multiplication to reduce the area, power consumption, and data storage required for real-time hardware implementation. This work also addresses the challenge of implementing the hardware to execute multiaxial contact-force estimation algorithms from a normal stress tactile sensor array on a field-programmable gate-array development platform, employing a high-level description approach. This paper describes the hardware implementation of the proposed sparse algorithm and that of an algorithm previously reported in the literature, comparing the results of both hardware implementations with the software results already validated. The calculation of force vectors on the proposed hardware required an average time of 58.68 ms, with an estimation error of 12.6% for normal forces and 7.7% for tangential forces on a 10 × 10 taxel tactile sensor array. Some advantages of the developed hardware are that it does not require additional memory elements, achieves a 4× reduction in processing elements compared to a non-sparse implementation, and meets the requirements of being generalizable, scalable, and efficient, allowing an expansion of the applications of normal stress sensors in low-power tactile systems. Full article

(This article belongs to the Special Issue Recent Development of Flexible Tactile Sensors and Their Applications)

► Show Figures

Figure 1

28 pages, 1950 KiB

Open AccessArticle

Type-2 Neutrosophic Markov Chain Model for Subject-Independent Sign Language Recognition: A New Uncertainty–Aware Soft Sensor Paradigm

by Muslem Al-Saidi, Áron Ballagi, Oday Ali Hassen and Saad M. Saad

Sensors 2024, 24(23), 7828; https://doi.org/10.3390/s24237828 (registering DOI) - 7 Dec 2024

Abstract

Uncertainty-aware soft sensors in sign language recognition (SLR) integrate methods to quantify and manage the uncertainty in their predictions. This is particularly crucial in SLR due to the variability in sign language gestures and differences in individual signing styles. Managing uncertainty allows the [...] Read more.

Uncertainty-aware soft sensors in sign language recognition (SLR) integrate methods to quantify and manage the uncertainty in their predictions. This is particularly crucial in SLR due to the variability in sign language gestures and differences in individual signing styles. Managing uncertainty allows the system to handle variations in signing styles, lighting conditions, and occlusions more effectively. While current techniques for handling uncertainty in SLR systems offer significant benefits in terms of improved accuracy and robustness, they also come with notable disadvantages. High computational complexity, data dependency, scalability issues, sensor and environmental limitations, and real-time constraints all pose significant hurdles. The aim of the work is to develop and evaluate a Type-2 Neutrosophic Hidden Markov Model (HMM) for SLR that leverages the advanced uncertainty handling capabilities of Type-2 neutrosophic sets. In the suggested soft sensor model, the Foot of Uncertainty (FOU) allows Type-2 Neutrosophic HMMs to represent uncertainty as intervals, capturing the range of possible values for truth, falsity, and indeterminacy. This is especially useful in SLR, where gestures can be ambiguous or imprecise. This enhances the model’s ability to manage complex uncertainties in sign language gestures and mitigate issues related to model drift. The FOU provides a measure of confidence for each recognition result by indicating the range of uncertainty. By effectively addressing uncertainty and enhancing subject independence, the model can be integrated into real-life applications, improving interactions, learning, and accessibility for the hearing-impaired. Examples such as assistive devices, educational tools, and customer service automation highlight its transformative potential. The experimental evaluation demonstrates the superiority of the Type-2 Neutrosophic HMM over the Type-1 Neutrosophic HMM in terms of accuracy for SLR. Specifically, the Type-2 Neutrosophic HMM consistently outperforms its Type-1 counterpart across various test scenarios, achieving an average accuracy improvement of 10%. Full article

(This article belongs to the Special Issue Computer Vision and Smart Sensors for Human-Computer Interaction)

13 pages, 2104 KiB

Open AccessArticle

Sensor-Based Frailty Assessment Using Fitbit

by Mohammad Hosseinalizadeh, Mehran Asghari and Nima Toosizadeh

Sensors 2024, 24(23), 7827; https://doi.org/10.3390/s24237827 (registering DOI) - 7 Dec 2024

Abstract

This study evaluated the reliability of Fitbit in assessing frailty based on motor and heart rate (HR) parameters through a validated upper extremity function (UEF) test, which involves 20 s of rapid elbow flexion. For motor performance, participants completed six trials of full [...] Read more.

This study evaluated the reliability of Fitbit in assessing frailty based on motor and heart rate (HR) parameters through a validated upper extremity function (UEF) test, which involves 20 s of rapid elbow flexion. For motor performance, participants completed six trials of full elbow flexion using their right arm, with and without weight. Fitbit and a commercial motion sensor were worn on the right arm. For HR measurements, an ECG system was placed on the left chest alongside the Fitbit on the left wrist. Motor parameters assessing speed, flexibility, weakness, exhaustion, and HR before, during, and after UEF were measured. A total of 42 participants (age = 22 ± 3) were recruited. For motor parameters, excellent agreement was observed between the wearable sensor and Fitbit, except for flexibility (ICC = 0.87 ± 0.09). For HR parameters, ICC values showed weak agreement between ECG and Fitbit for HR increase and recovery (ICC = 0.24 ± 0.11), while moderate to stronger agreement was seen for mean HR during baseline, task, and post-task (ICC = 0.81 ± 0.13). Fitbit is a reliable tool for assessing frailty through motor parameters and provides reasonably accurate HR estimates during baseline, task, and recovery periods. However, Fitbit’s ability to track rapid HR changes during activity is limited. Full article

(This article belongs to the Section Wearables)

► Show Figures

Figure 1

14 pages, 2080 KiB

Open AccessArticle

A XGBoost-Based Prediction Method for Meat Sheep Transport Stress Using Wearable Photoelectric Sensors and Infrared Thermometry

by Ruiqin Ma, Runqing Chen, Buwen Liang and Xinxing Li

Sensors 2024, 24(23), 7826; https://doi.org/10.3390/s24237826 (registering DOI) - 7 Dec 2024

Abstract

Transportation pressure poses a serious threat to the health of live sheep and the quality of their meat. So, the edible Hu sheep was chosen as the research object for meat sheep. We constructed a systematic biosignal detecting, processing, and modeling method. The [...] Read more.

Transportation pressure poses a serious threat to the health of live sheep and the quality of their meat. So, the edible Hu sheep was chosen as the research object for meat sheep. We constructed a systematic biosignal detecting, processing, and modeling method. The biosignal sensing was performed with wearable sensors (photoelectric sensor and infrared temperature measurement) for physiological dynamic sensing and continuous monitoring of the transport environment of meat sheep. Core waveform extraction and modern spectral estimation methods are used to determine and strip out the target signal waveform from it for the purpose of accurate sensing and the acquisition of key transport parameters. Subsequently, we built a qualitative stress assessment method based on external manifestations with reference to the Karolinska drowsiness scale to establish stage classification rules for monitoring data in the transportation environment of meat sheep. Finally, machine learning algorithms such as Gaussian Naive Bayes (GaussianNB), Passive-Aggressive Aggregative Classifier (PAC), Nearest Centroid (NC), K-Nearest Neighbor Classification (KNN), Random Forest (RF), Support Vector Classification (SVC), Gradient Boosting Decision Tree (GBDT), and eXtreme Gradient Boosting (XGB) were established to predict the classification models of transportation stress in meat sheep. Their classification results were compared. The results show that SVC and GBDT algorithms are more effective and the overall model classification accuracy reached 86.44% and 91.53%. XGB has the best results. The accuracy of the assessment of the transport stress state of meat sheep after the optimization of three parameters was 100%, 90.91%, and 93.33%, and the classification accuracy of the overall model reached 94.92%. The final results achieved improve transport reliability, reduce transport risk, and solve the problems of inefficient meat sheep transport supervision and quality control. Full article

(This article belongs to the Section Biosensors)

► Show Figures

Figure 1

12 pages, 1546 KiB

Open AccessArticle

Concurrent Validity and Relative Reliability of the RunScribe™ System for the Assessment of Spatiotemporal Gait Parameters During Walking

by Andrés Ráfales-Perucha, Elisa Bravo-Viñuales, Alejandro Molina-Molina, Antonio Cartón-Llorente, Silvia Cardiel-Sánchez and Luis E. Roche-Seruendo

Sensors 2024, 24(23), 7825; https://doi.org/10.3390/s24237825 (registering DOI) - 7 Dec 2024

Abstract

The evaluation of gait biomechanics using portable inertial measurement units (IMUs) offers real-time feedback and has become a crucial tool for detecting gait disorders. However, many of these devices have not yet been fully validated. The aim of this study was to assess [...] Read more.

The evaluation of gait biomechanics using portable inertial measurement units (IMUs) offers real-time feedback and has become a crucial tool for detecting gait disorders. However, many of these devices have not yet been fully validated. The aim of this study was to assess the concurrent validity and relative reliability of the RunScribe™ system for measuring spatiotemporal gait parameters during walking. A total of 460 participants (age: 36 ± 13 years; height: 173 ± 9 cm; body mass: 70 ± 13 kg) were asked to walk on a treadmill at 5 km·h⁻¹. Spatiotemporal parameters of step frequency (SF), step length (SL), step time (ST), contact time (CT), swing time (SwT), stride time (StT), stride length (StL) and normalized stride length (StL%) were measured through RunScribe™ and OptoGait™ systems. Bland–Altman analysis indicated small systematic biases and random errors for all variables. Pearson correlation analysis showed strong correlations (0.70–0.94) between systems. The intraclass correlation coefficient supports these results, except for contact time (ICC = 0.64) and swing time (ICC = 0.34). The paired t-test showed small differences in SL, StL and StL% (≤0.25) and large in CT and SwT (1.2 and 2.2, respectively), with no differences for the rest of the variables. This study confirms the accuracy of the RunScribe™ system for assessing spatiotemporal parameters during walking, potentially reducing the barriers to continuous gait monitoring and early detection of gait issues. Full article

(This article belongs to the Special Issue IMU and Innovative Sensors for Healthcare)

► Show Figures

Figure 1

20 pages, 8275 KiB

Open AccessArticle

Automated Visual Inspection for Precise Defect Detection and Classification in CBN Inserts

by Li Zeng, Feng Wan, Baiyun Zhang and Xu Zhu

Sensors 2024, 24(23), 7824; https://doi.org/10.3390/s24237824 (registering DOI) - 7 Dec 2024

Abstract

In the high-stakes domain of precision manufacturing, Cubic Boron Nitride (CBN) inserts are pivotal for their hardness and durability. However, post-production surface defects on these inserts can compromise product integrity and performance. This paper proposes an automated detection and classification system using machine [...] Read more.

In the high-stakes domain of precision manufacturing, Cubic Boron Nitride (CBN) inserts are pivotal for their hardness and durability. However, post-production surface defects on these inserts can compromise product integrity and performance. This paper proposes an automated detection and classification system using machine vision to scrutinize these surface defects. By integrating an optical bracket, a high-resolution industrial camera, precise lighting, and an advanced development board, the system employs digital image processing to ascertain and categorize imperfections on CBN inserts. The methodology initiates with a high-definition image capture by the imaging platform, tailored for CBN insert inspection. A suite of defect detection algorithms undergoes comparative analysis to discern their efficacy, emphasizing the impact of algorithm parameters and dataset diversity on detection precision. The most effective algorithm is then encapsulated into a versatile application, ensuring compatibility with various operating systems. Empirical verification of the system shows that the detection accuracy of multiple defect types exceeds 90%, and the tooth surface recognition efficiency significantly reaches three frames per second, with the front and side cutting surfaces of the tool in each frame. This breakthrough indicates a scalable, reliable solution for automatically detecting and classifying surface defects on CBN inserts, paving the way for enhanced quality control in automated, high-speed production lines. Full article

(This article belongs to the Special Issue Dalian University of Technology Celebrating 75th Anniversary)

► Show Figures

Figure 1

12 pages, 6318 KiB

Open AccessArticle

An Ultrasensitive Ethanol Gas Sensor Based on a Dual-Nanoparticle In₂O₃/SnO₂ Composite

by Cheng Zhang, Ze Zhang, Yao Tian, Lingmin Yu and Hairong Wang

Sensors 2024, 24(23), 7823; https://doi.org/10.3390/s24237823 (registering DOI) - 7 Dec 2024

Abstract

As a VOC, ethanol can be found in human exhaled breath, and its concentration can be used as a biomarker of human liver disease. To detect trace-level concentrations of ethanol, an ultrasensitive ethanol sensor was developed based on a dual-nanoparticle In₂O [...] Read more.

As a VOC, ethanol can be found in human exhaled breath, and its concentration can be used as a biomarker of human liver disease. To detect trace-level concentrations of ethanol, an ultrasensitive ethanol sensor was developed based on a dual-nanoparticle In₂O₃/SnO₂ composite that was prepared by hydrothermal synthesis, and its suspension was dipped on a flat electrode to form a gas sensor. The nanocomposite was characterized by an SEM (scanning electron microscope), XRD (X-ray diffraction), and a TEM (transmission electron microscope), and the nanoparticle structure was observed. The experimental results showed that gas sensors based on the In₂O₃/SnO₂ nanocomposite had higher responses compared to sensors based on pure In₂O₃. Among the nanocomposites, the one with a In₂O₃-to-SnO₂ mol ratio of 1:8 was used in the sensor with the highest response of 1.41 to 100 ppb ethanol at 150 °C, which also exhibited good repeatability. The ultrasensitive response to ethanol can be attributed to the faster electron migration rate and the increase in oxygen-absorbing sites caused by the n-n heterojunction in the nanocomposite. Due to its low detection limit, good repeatability, and relatively high responses in high humidity, this sensor has a potential application in exhaled breath detection. Full article

(This article belongs to the Section Chemical Sensors)

► Show Figures

Figure 1

18 pages, 30948 KiB

Open AccessArticle

Improving Circulating Tumor Cell Detection Using Image Synthesis and Transformer Models in Cancer Diagnostics

by Shuang Liang, Xue Bai and Yu Gu

Sensors 2024, 24(23), 7822; https://doi.org/10.3390/s24237822 (registering DOI) - 7 Dec 2024

Abstract

Cancer is the second leading cause of death, significantly threatening human health. Effective treatment options are often lacking in advanced stages, making early diagnosis crucial for reducing mortality rates. Circulating tumor cells (CTCs) are a promising biomarker for early detection; however, their automatic [...] Read more.

Cancer is the second leading cause of death, significantly threatening human health. Effective treatment options are often lacking in advanced stages, making early diagnosis crucial for reducing mortality rates. Circulating tumor cells (CTCs) are a promising biomarker for early detection; however, their automatic detection is challenging due to their heterogeneous size and shape, as well as their scarcity in blood. This study proposes a data generation method using the Segment Anything Model (SAM) combined with a copy–paste strategy. We develop a detection network based on the Swin Transformer, featuring a backbone network, scale adapter module, shape adapter module, and detection head, which enhances CTC localization and identification in images. To effectively utilize both generated and real data, we introduce an improved loss function that includes a regularization term to ensure consistency across different data distributions. Our model demonstrates exceptional performance across five evaluation metrics: accuracy (0.9960), recall (0.9961), precision (0.9804), specificity (0.9975), and mean average precision (mAP) of 0.9400 at an Intersection over Union (IoU) threshold of 0.5. These results are achieved on a dataset generated by mixing both public and local data, highlighting the robustness and generalizability of the proposed approach. This framework surpasses state-of-the-art models (ADCTC, DiffusionDet, CO-DETR, and DDQ), providing a vital tool for early cancer diagnosis, treatment planning, and prognostic assessment, ultimately enhancing human health and well-being. Full article

(This article belongs to the Special Issue Image Analysis and Biomedical Sensors)

► Show Figures

Figure 1

20 pages, 3564 KiB

Open AccessArticle

MOMFNet: A Deep Learning Approach for InSAR Phase Filtering Based on Multi-Objective Multi-Kernel Feature Extraction

by Xuedong Zhang, Cheng Peng, Ziqi Li, Yaqi Zhang, Yongxuan Liu and Yong Wang

Sensors 2024, 24(23), 7821; https://doi.org/10.3390/s24237821 - 6 Dec 2024

Abstract

Interferometric Synthetic Aperture Radar (InSAR) is a widely used remote sensing technology for Earth observation, enabling the detection and measurement of ground deformation through the generation of interferograms. However, phase noise remains a critical factor that degrades interferogram quality. To address this issue, [...] Read more.

Interferometric Synthetic Aperture Radar (InSAR) is a widely used remote sensing technology for Earth observation, enabling the detection and measurement of ground deformation through the generation of interferograms. However, phase noise remains a critical factor that degrades interferogram quality. To address this issue, this study proposes MOMFNet, a deep learning approach for InSAR phase filtering based on multi-objective multi-kernel feature extraction that leverages multi-objective multi-kernel feature extraction. MOMFNet incorporates a multi-objective loss function that accounts for both the spatial and statistical characteristics of the denoising results, while its multi-kernel convolutional feature extraction module captures multi-scale information comprehensively. Furthermore, the introduction of weighted residual blocks allows the model to adaptively adjust the importance of features, improving its ability to accurately identify and suppress noise. To train the MOMFNet network, we developed an interferogram simulation strategy that uses randomly distorted 2D Gaussian surfaces to simulate terrain variations, Perlin noise to model atmospheric turbulence phases, and negative Gaussian noise to generate random training samples at different noise levels. Comparative experiments with traditional denoising methods and other deep learning approaches, through both qualitative and quantitative analyses, demonstrated that MOMFNet excels in noise suppression and phase recovery, particularly in scenarios involving large gradients and random noise. Empirical studies using Sentinel-1 satellite data from the Yanzhou coal mine validated the practical value of MOMFNet, showing that it effectively removes irrelevant noise while preserving critical phase details, significantly improving interferogram quality. This research provides important insights into the application of deep learning for InSAR denoising. Full article

(This article belongs to the Section Radar Sensors)

17 pages, 6741 KiB

Open AccessArticle

Design and Monitoring Application of an Adjustable Intelligent Bearing Based on Pressure Sensing

by Shu Li, Zaiyu Zhang, Luyi Gan, Jiheng Yin and Ming Fu

Sensors 2024, 24(23), 7820; https://doi.org/10.3390/s24237820 - 6 Dec 2024

Abstract

Single-pier, dual-bearing bridges are susceptible to effects such as concrete creep, thermal expansion, and uneven foundation settlement. When combined with eccentric loading from heavy vehicles, these factors collectively can significantly increase the risk of bridge overturning. To address this risk, a comprehensive analysis [...] Read more.

Single-pier, dual-bearing bridges are susceptible to effects such as concrete creep, thermal expansion, and uneven foundation settlement. When combined with eccentric loading from heavy vehicles, these factors collectively can significantly increase the risk of bridge overturning. To address this risk, a comprehensive analysis of the bridge overturning mechanism was conducted. Considering the current limitations of health monitoring in bearing reaction force (BRF) measurement and risk mitigation, an adjustable intelligent bearing based on pressure sensing and self-locking principles was developed. Its mechanical performance was analyzed under the most unfavorable load conditions. To further validate the approach, a specific experimental bridge was used as a case study. The effectiveness of the force measurement and height adjustment functions was evaluated through moving load experiments. The results showed that the force measurement function accurately captured dynamic BRF changes within a precision range of ±0.1% FS and demonstrated high sensitivity to instantaneous impact effects. The height adjustment function achieved a reaction force change of up to 40 kN within the maximum adjustment range of 1.2 mm, significantly improving the load distribution of the bridge. These findings validated the reliability of the proposed intelligent bearing in real-time monitoring and proactive risk adjustment. This effectively overcomes the limitations of existing bearings, which only perform passive monitoring. Overall, it achieves the real-time monitoring of BRF and proactive control of bridge overturning risks. Full article

(This article belongs to the Special Issue Sensors for Structural Health Monitoring of Structures and Infrastructures)

► Show Figures

Figure 1

12 pages, 992 KiB

Open AccessArticle

Steps to Facilitate the Use of Clinical Gait Analysis in Stroke Patients: The Validation of a Single 2D RGB Smartphone Video-Based System for Gait Analysis

by Philipp Barzyk, Alina-Sophie Boden, Justin Howaldt, Jana Stürner, Philip Zimmermann, Daniel Seebacher, Joachim Liepert, Manuel Stein, Markus Gruber and Michael Schwenk

Sensors 2024, 24(23), 7819; https://doi.org/10.3390/s24237819 - 6 Dec 2024

Abstract

Clinical gait analysis plays a central role in the rehabilitation of stroke patients. However, practical and technical challenges limit their use in clinical settings. This study aimed to validate SMARTGAIT, a deep learning-based gait analysis system that addresses these limitations. Eight stroke patients [...] Read more.

Clinical gait analysis plays a central role in the rehabilitation of stroke patients. However, practical and technical challenges limit their use in clinical settings. This study aimed to validate SMARTGAIT, a deep learning-based gait analysis system that addresses these limitations. Eight stroke patients took part in the study at the Human Performance Research Centre of the University of Konstanz. Gait measurements were taken using both the marker-based Vicon motion capture system and the single-smartphone-based SMARTGAIT system. We evaluated the agreement for knee, hip, and ankle joint angle kinematics in the frontal and sagittal plane and spatiotemporal gait parameters between the two systems. The results mostly demonstrated high levels of agreement between the two systems, with Pearson correlations of ≥0.79 for all lower body angle kinematics in the sagittal plane and correlations of ≥0.71 in the frontal plane. RMSE values were ≤4.6°. The intraclass correlation coefficients for all derived gait parameters showed good to excellent levels of agreement. SMARTGAIT is a promising tool for gait analysis in stroke, particularly for quantifying gait characteristics in the sagittal plane, which is very relevant for clinical gait analysis. However, further analyses are required to validate the use of SMARTGAIT in larger samples and its transferability to different types of pathological gait. In conclusion, a single smartphone recording (monocular 2D RGB camera) could make gait analysis more accessible in clinical settings, potentially simplifying the process and making it more feasible for therapists and doctors to use in their day-to-day practice. Full article

(This article belongs to the Section Intelligent Sensors)

► Show Figures

Figure 1

24 pages, 5193 KiB

Open AccessArticle

Optimising Concrete Crack Detection: A Study of Transfer Learning with Application on Nvidia Jetson Nano

by C. Long Nguyen, Andy Nguyen, Jason Brown, Terry Byrne, Binh Thanh Ngo and Chieu Xuan Luong

Sensors 2024, 24(23), 7818; https://doi.org/10.3390/s24237818 - 6 Dec 2024

Abstract

The use of Artificial Intelligence (AI) to detect defects such as concrete cracks in civil and transport infrastructure has the potential to make inspections less expensive, quicker, safer and more objective by reducing the need for on-site human labour. One deployment scenario involves [...] Read more.

The use of Artificial Intelligence (AI) to detect defects such as concrete cracks in civil and transport infrastructure has the potential to make inspections less expensive, quicker, safer and more objective by reducing the need for on-site human labour. One deployment scenario involves using a drone to carry an embedded device and camera, with the device making localised predictions at the edge about the existence of defects using a trained convolutional neural network (CNN) for image classification. In this paper, we trained six CNNs, namely Resnet18, Resnet50, GoogLeNet, MobileNetV2, MobileNetV3-Small and MobileNetV3-Large, using transfer learning technology to classify images of concrete structures as containing a crack or not. To enhance the model’s robustness, the original dataset, comprising 3000 images of concrete structures, was augmented using salt and pepper noise, as well as motion blur, separately. The results show that Resnet50 generally provides the highest validation accuracy (96% with the original dataset and a batch size of 16) and the highest validation F1-score (95% with the original dataset and a batch size of 16). The trained model was then deployed on an Nvidia Jetson Nano device for real-time inference, demonstrating its capability to accurately detect cracks in both laboratory and field settings. This study highlights the potential of using transfer learning on Edge AI devices for Structural Health Monitoring, providing a cost-effective and efficient solution for automated crack detection in concrete structures. Full article

(This article belongs to the Special Issue Smart Sensors for Transportation Infrastructure Health Monitoring)

Journal Description

Sensors

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI