Showing papers in "Journal of Sensors in 2019"

Detection of Apple Lesions in Orchards Based on Deep Learning Methods of CycleGAN and YOLOV3-Dense

Abstract: Plant disease is one of the primary causes of crop yield reduction. With the development of computer vision and deep learning technology, autonomous detection of plant surface lesion images collected by optical sensors has become an important research direction for timely crop disease diagnosis. In this paper, an anthracnose lesion detection method based on deep learning is proposed. Firstly, for the problem of insufficient image data caused by the random occurrence of apple diseases, in addition to traditional image augmentation techniques, Cycle-Consistent Adversarial Network (CycleGAN) deep learning model is used in this paper to accomplish data augmentation. These methods effectively enrich the diversity of training data and provide a solid foundation for training the detection model. In this paper, on the basis of image data augmentation, densely connected neural network (DenseNet) is utilized to optimize feature layers of the YOLO-V3 model which have lower resolution. DenseNet greatly improves the utilization of features in the neural network and enhances the detection result of the YOLO-V3 model. It is verified in experiments that the improved model exceeds Faster R-CNN with VGG16 NET, the original YOLO-V3 model, and other three state-of-the-art networks in detection performance, and it can realize real-time detection. The proposed method can be well applied to the detection of anthracnose lesions on apple surfaces in orchards.

Fundamentals, Recent Advances, and Future Challenges in Bioimpedance Devices for Healthcare Applications

Abstract: This work develops a thorough review of bioimpedance systems for healthcare applications. The basis and fundamentals of bioimpedance measurements are described covering issues ranging from the hardware diagrams to the configurations and designs of the electrodes and from the mathematical models that describe the frequency behavior of the bioimpedance to the sources of noise and artifacts. Bioimpedance applications such as body composition assessment, impedance cardiography (ICG), transthoracic impedance pneumography, electrical impedance tomography (EIT), and skin conductance are described and analyzed. A breakdown of recent advances and future challenges of bioimpedance is also performed, addressing topics such as transducers for biosensors and Lab-on-Chip technology, measurements in implantable systems, characterization of new parameters and substances, and novel bioimpedance applications.

Sensors of Smart Devices in the Internet of Everything (IoE) Era: Big Opportunities and Massive Doubts

Abstract: Smart device industry allows developers and designers to embed different sensors, processors, and memories in small-size electronic devices. Sensors are added to enhance the usability of these devices and improve the quality of experience through data collection and analysis. However, with the era of big data and machine learning, sensors’ data may be processed by different techniques to infer various hidden information. The extracted information may be beneficial to device users, developers, and designers to enhance the management, operation, and development of these devices. However, the extracted information may be used to compromise the security and the privacy of humans in the era of Internet of Everything (IoE). In this work, we attempt to review the process of inferring meaningful data from smart devices’ sensors, especially, smartphones. In addition, different useful machine learning applications based on smartphones’ sensors data are shown. Moreover, different side channel attacks utilizing the same sensors and the same machine learning algorithms are overviewed.

A Benchmarking of Learning Strategies for Pest Detection and Identification on Tomato Plants for Autonomous Scouting Robots Using Internal Databases

Abstract: Greenhouse crop production is growing throughout the world and early pest detection is of particular importance in terms of productivity and reduction of the use of pesticides. Conventional eye observation methods are nonefficient for large crops. Computer vision and recent advances in deep learning can play an important role in increasing the reliability and productivity. This paper presents the development and comparison of two different approaches for vision based automated pest detection and identification, using learning strategies. A solution that combines computer vision and machine learning is compared against a deep learning solution. The main focus of our work is on the selection of the best approach based on pest detection and identification accuracy. The inspection is focused on the most harmful pests on greenhouse tomato and pepper crops, Bemisia tabaci and Trialeurodes vaporariorum. A dataset with a huge number of infected tomato plants images was created to generate and evaluate machine learning and deep learning models. The results showed that the deep learning technique provides a better solution because (a) it achieves the disease detection and classification in one step, (b) gets better accuracy, (c) can distinguish better between Bemisia tabaci and Trialeurodes vaporariorum, and (d) allows balancing between speed and accuracy by choosing different models.

Traffic Flow Prediction during the Holidays Based on DFT and SVR

Abstract: With the implementation of the freeway free policy during the holidays, traffic congestion in the freeway becomes a common phenomenon. In order to alleviate traffic pressure, traffic flow prediction during the holidays has become a problem of great concern. This paper proposes a hybrid prediction methodology combining discrete Fourier transform (DFT) with support vector regression (SVR). The common trend in the traffic flow data is extracted using DFT by setting an appropriate threshold, which is predicted by extreme extrapolation of the historical trend. The SVR method is applied to predict the residual series. The experimental results with measured data collected from the toll stations in Jiangsu province of China show that the proposed algorithm has higher accuracy compared with the traditional method, and it is an efficient method for traffic flow prediction during the holidays.

Ion-Imprinted Electrochemical Sensor Based on Copper Nanoparticles-Polyaniline Matrix for Nitrate Detection

Abstract: This study reports a new chemical sensor based on ion-imprinted polymer matrix using copper nanoparticles-polyaniline nanocomposite (IIP-Cu-NPs/PANI). This sensor was prepared by electropolymerization using aniline as a functional monomer and nitrate as template onto the copper nanoparticles-modified glassy carbon (GC) electrode surface. Both ion-imprinted (IIP) and nonimprinted (NIP) electrochemical sensor surfaces were evaluated using UV-Visible spectrometry and scanning electron microscopy (SEM). The electrochemical analysis was made via cyclic voltammetry (CV), linear sweep voltammetry (LSV), and impedance spectroscopy (IS). Throughout this study various analytical parameters, such as scan rate, pH value, concentration of monomer and template, and electropolymerization cycles, were optimized. Under the optimum conditions, the peaks current of nitrate was linear to its concentration in the range of 1μM-0.1M with a detection limit of 31μM and 5μM by EIS and LSV. The developed imprinted nitrate sensor was successfully applied for nitrate determination in different real water samples with acceptable recovery rates.

WBSN in IoT Health-Based Application: Toward Delay and Energy Consumption Minimization

Abstract: The wireless body sensor network (WBSN) technologies are one of the essential technologies of the Internet of things (IoT) growths of the healthcare paradigm, where every patient is monitored through a group of small-powered and lightweight sensor nodes. Thus, energy consumption is a major issue in WBSN. The major causes of energy wastage in WBSN are collisions and retransmission process. However, the major cause of the collision happened when two sensors are attempting to transmit data at exactly the same time and same frequency, and the major cause of the retransmission process happened when the collision takes place or data does not received properly due to channel fading. In this paper, we proposed a cognitive cooperative communication with two master nodes, namely, as two cognitive master nodes (TCMN), which can eliminate the collision and reduce the retransmission process. First, a complete study of a scheme is investigated in terms of network architecture. Second, a mathematical model of the link and outage probability of the proposed protocol are derived. Third, the end-to-end delay, throughput, and energy consumption are analyzed and investigated. The simulation and numerical results show that the TCMN can do system performance under general conditions with respect to direct transmission mode (DTM) and existing work.

A Deep Learning Model for Concrete Dam Deformation Prediction Based on RS-LSTM

Abstract: Deformation is a comprehensive reflection of the structural state of a concrete dam, and research on prediction models for concrete dam deformation provides the basis for safety monitoring and early warning strategies. This paper focuses on practical problems such as multicollinearity among factors; the subjectivity of factor selection; robustness, externality, generalization, and integrity deficiencies; and the unsoundness of evaluation systems for prediction models. Based on rough set (RS) theory and a long short-term memory (LSTM) network, single-point and multipoint concrete dam deformation prediction models for health monitoring based on RS-LSTM are studied. Moreover, a new prediction model evaluation system is proposed, and the model accuracy, robustness, externality, and generalization are defined as quantitative evaluation indexes. An engineering project shows that the concrete dam deformation prediction models based on RS-LSTM can quantitatively obtain the representative factors that affect dam deformation and the importance of each factor relative to the effect. The accuracy evaluation index (AVI), robustness evaluation index (RVI), externality evaluation index (EVI), and generalization evaluation index (GVI) of the model are superior to the evaluation indexes of existing shallow neural network models and statistical models according to the new evaluation system, which can estimate the comprehensive performance of prediction models. The prediction model for concrete dam deformation based on RS-LSTM optimizes the factors that influence the model, quantitatively determines the importance of each factor, and provides high-performance, synchronous, and dynamic predictions for concrete dam behaviours; therefore, the model has strong engineering practicality.

Review of Wireless Acoustic Sensor Networks for Environmental Noise Monitoring in Smart Cities

Abstract: Nowadays, more than half of the world’s population lives in urban areas. Since this proportion is expected to keep rising, the sustainable development of cities is of paramount importance to guarantee the quality of life of their inhabitants. Environmental noise is one of the main concerns that has to be addressed, due to its negative impact on the health of people. Different national and international noise directives and legislations have been defined during the past decades, which local authorities must comply with involving noise mapping, action plans, policing, and public awareness, among others. To this aim, a recent change in the paradigm for environmental noise monitoring has been driven by the rise of Internet of Things technology within smart cities through the design and development of wireless acoustic sensor networks (WASNs). This work reviews the most relevant WASN-based approaches developed to date focused on environmental noise monitoring. The proposals have moved from networks composed of high-accuracy commercial devices to the those integrated by ad hoc low-cost acoustic sensors, sometimes designed as hybrid networks with low and high computational capacity nodes. After describing the main characteristics of recent WASN-based projects, the paper also discusses several open challenges, such as the development of acoustic signal processing techniques to identify noise events, to allow the reliable and pervasive deployment of WASNs in urban areas together with some potential future applications.

Design Approaches of MEMS Microphones for Enhanced Performance

Abstract: This paper reports a review about microelectromechanical system (MEMS) microphones. The focus of this review is to identify the issues in MEMS microphone designs and thoroughly discuss the state-of-the-art solutions that have been presented by the researchers to improve performance. Considerable research work has been carried out in capacitive MEMS microphones, and this field has attracted the research community because these designs have high sensitivity, flat frequency response, and low noise level. A detailed overview of the omnidirectional microphones used in the applications of an audio frequency range has been presented. Since the microphone membrane is made of a thin film, it has residual stress that degrades the microphone performance. An in-depth detailed review of research articles containing solutions to relieve these stresses has been presented. The comparative analysis of fabrication processes of single- and dual-chip omnidirectional microphones, in which the membranes are made up of single-crystal silicon, polysilicon, and silicon nitride, has been done, and articles containing the improved performance in these two fabrication processes have been explained. This review will serve as a starting guide for new researchers in the field of capacitive MEMS microphones.

Intrusion Detection in Wireless Sensor Networks with an Improved NSA Based on Space Division

Abstract: Inspired by the biological immune system, many researchers apply artificial immune principles to intrusion detection in wireless sensor networks, such as negative selection algorithms, danger theory, and dendritic cell algorithms. When applying the negative selection algorithm to wireless sensor networks, the characteristics of wireless sensor networks, such as frequent changes in network topology and limited resources, are not considered too much, which makes the detection effect to need improvement. In this paper, a negative selection algorithm based on spatial partition is proposed and applied to hierarchical wireless sensor networks. The algorithm first analyzes the distribution of self-set in the real-valued space then divides the real-valued space, and several subspaces are obtained. Selves are filled into different subspaces. We implement the negative selection algorithm in the subspace. The randomly generated candidate detector only needs to be tolerated with selves in the subspace where the detector is located, not all the selves. This operation reduces the time cost of distance calculation. In the detection process of detectors, the antigen which is to be detected only needs to match the mature detectors in the subspace where the antigen is located, rather than all the detectors. This operation speeds up the antigen detection process. Theoretical analysis and experimental results show that the algorithm has better time efficiency and quality of detectors, saves sensor node resources and reduces the energy consumption, and is an effective algorithm for wireless sensor network intrusion detection.

Traffic Sensing Methodology Combining Influence Line Theory and Computer Vision Techniques for Girder Bridges

Abstract: Collecting the information of traffic load, especially heavy trucks, is crucial for bridge statistical analysis, safety evaluation, and maintenance strategies. This paper presents a traffic sensing methodology that combines a deep learning based computer vision technique with the influence line theory. Theoretical background and derivations are introduced from both aspects of structural analysis and computer vision techniques. In addition, to evaluate the effectiveness and accuracy of the proposed traffic sensing method through field tests, a systematic analysis is performed on a continuous box-girder bridge. The obtained results show that the proposed method can automatically identify the vehicle load and speed with promising efficiency and accuracy and most importantly cost-effectiveness. All these features make the proposed methodology a desirable bridge weigh-in-motion system, especially for bridges already equipped with structural health monitoring system.

Real-Time Vehicle Detection Algorithm Based on Vision and Lidar Point Cloud Fusion

Abstract: Vehicle detection is one of the most important environment perception tasks for autonomous vehicles. The traditional vision-based vehicle detection methods are not accurate enough especially for small and occluded targets, while the light detection and ranging- (lidar-) based methods are good in detecting obstacles but they are time-consuming and have a low classification rate for different target types. Focusing on these shortcomings to make the full use of the advantages of the depth information of lidar and the obstacle classification ability of vision, this work proposes a real-time vehicle detection algorithm which fuses vision and lidar point cloud information. Firstly, the obstacles are detected by the grid projection method using the lidar point cloud information. Then, the obstacles are mapped to the image to get several separated regions of interest (ROIs). After that, the ROIs are expanded based on the dynamic threshold and merged to generate the final ROI. Finally, a deep learning method named You Only Look Once (YOLO) is applied on the ROI to detect vehicles. The experimental results on the KITTI dataset demonstrate that the proposed algorithm has high detection accuracy and good real-time performance. Compared with the detection method based only on the YOLO deep learning, the mean average precision (mAP) is increased by 17%.

Research on Classification Method of Maize Seed Defect Based on Machine Vision

Abstract: Traditionally, the classification of seed defects mainly relies on the characteristics of color, shape, and texture. This method requires repeated extraction of a large amount of feature information, which is not efficiently used in detection. In recent years, deep learning has performed well in the field of image recognition. We introduced convolutional neural networks (CNNs) and transfer learning into the quality classification of seeds and compared them with traditional machine learning algorithms. Experiments showed that deep learning algorithm was significantly better than the machine learning algorithm with an accuracy of 95% (GoogLeNet) vs. 79.2% (SURF+SVM). We used three classifiers in GoogLeNet to demonstrate that network accuracy increases as the depth of the network increases. We used the visualization technology to obtain the feature map of each layer of the network in CNNs and used the heat map to represent the probability distribution of the inference results. As an end-to-end network, CNNs can be easily applied for automated seed manufacturing.

A Low-Cost Closed-Loop Solar Tracking System Based on the Sun Position Algorithm

Abstract: Sun position and the optimum inclination of a solar panel to the sun vary over time throughout the day. A simple but accurate solar position measurement system is essential for maximizing the output power from a solar panel in order to increase the panel efficiency while minimizing the system cost. Solar position can be measured either by a sensor (active/passive) or through the sun position monitoring algorithm. Sensor-based sun position measuring systems fail to measure the solar position in a cloudy or intermittent day, and they require precise installation and periodic calibrations. In contrast, the sun position algorithms use mathematical formula or astronomical data to obtain the station of the sun at a particular geographical location and time. A standalone low-cost but high-precision dual-axis closed-loop sun-tracking system using the sun position algorithm was implemented in an 8-bit microcontroller platform. The Astronomical Almanac’s (AA) algorithm was used for its simplicity, reliability, and fast computation capability of the solar position. Results revealed that incorporation of the sun position algorithm into a solar tracking system helps in outperforming the fixed system and optical tracking system by 13.9% and 2.1%, respectively. In summary, even for a small-scale solar tracking system, the algorithm-based closed-loop dual-axis tracking system can increase overall system efficiency.

LoRa (Long-Range) High-Density Sensors for Internet of Things

Abstract: Over the past few years, there has been a growing awareness regarding the concept of Internet of Things (IoT), which involves connecting to the Internet various objects surrounding us in everyday life. The main purpose of this concept closely connected to the smart city issue is increasing the quality of life by contributing to streamlining resource consumption and protecting the environment. The LoRa communication mechanism is a physical layer of the LoRaWAN protocol, defined by the LoRa Alliance. Compared to other existing technologies, LoRa is a modulation technique enabling the transfer of information over a range of tens of kilometers. The main contribution this paper brings to the field is analyzing the scalability of the LoRa technology and determining the maximum number of sensors which can be integrated into this type of monitoring and control architecture. The sensor architecture is specific to the smart city concept that involves the integration of a large number of high-density sensors distributed on a large-scale geographic area. The reason behind this study is the need to assess the scalability of the LoRa technology, taking into consideration other factors, such as the packet payload size, the duty circle parameter, the spreading factor, and the number of nodes. The experimental results reveal that the maximum number of LoRa sensors that can communicate on the same channel is 1,500; furthermore, in order to obtain a high performance level, it is necessary to schedule and plan the network as carefully as possible. The spreading factor must be allocated according to the distance at which the sensor is placed from the gateway.

A Survey on Pollution Monitoring Using Sensor Networks in Environment Protection

Abstract: Detecting pollution timely and locating the pollution source is of great importance in environmental protection. Considering advantages of the sensor network technology, sensor networks have been adopted in pollution monitoring works. In this paper, a survey on researches of pollution monitoring using sensor networks in environment protection is given. Firstly, sensors and pollution monitoring network systems are studied. Secondly, different pollution detection methods are analyzed and compared. Thirdly, an overview of state-of-art technologies on pollution source localization is given. Finally, challenges on pollution monitoring using sensor networks are presented.

Enhanced Sensitivity of Surface Plasmon Resonance Biosensor Functionalized with Doped Polyaniline Composites for the Detection of Low-Concentration Acetone Vapour

Abstract: PANI/chitosan composite and a ternary composite comprising of PANI, chitosan, and reduced graphene oxide have been successfully synthesised and characterised using FTIR and UV-VIS spectroscopy. Optical constants of the composites were extracted from the UV-VIS spectra. The extracted parameters were applied in the simulation of a surface plasmon resonance (SPR) biosensor functionalised with PANI/chitosan and ternary composites. The aim was to explore the applicability of the composite-based SPR sensor in the detection of low-concentration acetone vapour within the range of 1.8 ppm–5.0 ppm for diabetes monitoring and screening. The functionalization of the SPR sensor with the PANI/chitosan and the ternary composites shows promising application of the sensor in the detection of acetone vapour at a low concentration down to less than 0.5 ppm. The maximum sensitivity values of about 60 and 180 degree/refractive index change were observed for PANI/chitosan and ternary composite sensing layers, respectively, in comparison with the bare gold-based SPR which shows no response up to 10 ppm concentration of acetone vapour in air. In addition, the two sensing layers show good selectivity to acetone vapour compared to ethanol, methanol, and ammonia. The response in the case of ternary composite shows better linearity with a correlation coefficient of 1.0 compared to PANI/chitosan- and gold-based SPR layers with 0.9999 and 0.9997, respectively.

Detection of Ascorbic Acid Using Green Synthesized Carbon Quantum Dots

Abstract: In this work, carbon quantum dots (CQDs) were synthesized by microwave irradiation and were electropolymerized on glassy carbon electrode (GCE) to establish an electrochemical sensor for the selective detection of ascorbic acid (AA). Electrochemical behaviors of the prepared sensor were investigated by cyclic voltammetry (CV), differential pulse voltammetry (DPV), and electrochemical impedance spectroscopy (EIS). Herein, two wide linear responses were obtained in ranges of 0.01-3 mM and 4-12 mM with a low detection limit of 10 μM to AA. High sensitivities (44.13 μA-1 μM-1 cm-2, 9.66 μA-1 μM-1 cm-2, respectively) corresponding to the linear ranges were also achieved. In addition, the electrochemical sensor exhibited good selectivity and robust anti-interference ability toward AA in the presence of dopamine (DA) and uric acid (UA). These results showed that this sensor can be used as a promising tool to detect AA in real complex systems.

Biometric-Based Attendance Tracking System for Education Sectors: A Literature Survey on Hardware Requirements

Abstract: The application of biometric recognition in personal authentication enables the growth of this technology to be employed in various domains. The implementation of biometric recognition systems can be based on physical or behavioral characteristics, such as the iris, voice, fingerprint, and face. Currently, the attendance tracking system based on biometric recognition for education sectors is still underutilized, thus providing a good opportunity to carry out interesting research in this area. As evidenced in a typical classroom, educators tend to take the attendance of their students by using conventional methods such as by calling out names or signing off an attendance sheet. Yet, these types of methods are proved to be time consuming and tedious, and sometimes, fraud occurs. As a result, significant progress had been made to mark attendance automatically by making use of biometric recognition. This progress enables a new and more advanced biometric-based attendance system being developed over the past ten years. The setting-up of biometric-based attendance systems requires both software and hardware components. Since the software and hardware sections are too broad to be discussed in one paper, this literature survey only provides an overview of the types of hardware used. Emphasis is then placed on the microcontroller platform, biometric sensor, communication channel, database storage, and other components in order to assist future researchers in designing the hardware part of biometric-based attendance systems.

DV-Hop Node Location Algorithm Based on GSO in Wireless Sensor Networks

Abstract: Node location is one of the most important problems to be solved in practical application of WSN. As a typical location algorithm without ranging, DV-Hop is widely used in node localization of wireless sensor networks. However, in the third phase of DV-Hop, a least square method is used to solve the nonlinear equations. Using this method to locate the unknown nodes will produce large coordinate errors, poor stability of positioning accuracy, low location coverage, and high energy consumption. An improved localization algorithm based on hybrid chaotic strategy (MGDV-Hop) is proposed in this paper. Firstly, a glowworm swarm optimization of hybrid chaotic strategy based on chaotic mutation and chaotic inertial weight updating (MC-GSO) is proposed. The MC-GSO algorithm is used to control the moving distance of each firefly by chaos mutation and chaotic inertial weight when the firefly falls into a local optimum. The experimental results show that MC-GSO has better convergence and higher accuracy and avoids the premature convergence. Then, MC-GSO is used to replace the least square method in estimating node coordinates to solve the problem that the localization accuracy of the DV-Hop algorithm is not high. By establishing the error fitness function, the linear solution of coordinates is transformed into a two-dimensional combinatorial optimization problem. The simulation results and analysis confirm that the improved algorithm (MGDV-Hop) reduces the average location error, increases the location coverage, and decreases and balances the energy consumption as compared to DV-Hop and the location algorithm based on classical GSO (GSDV-Hop).

Construction of Indoor Thermal Comfort Environmental Monitoring System Based on the IoT Architecture

Abstract: With the development and progress of technology, people’s requirements for living quality are increasingly higher. This study builds an indoor thermal comfort environmental monitoring system through the Internet of Things (IoT) architecture to explore the thermal comfort of people in indoor environments. Then, the applicable indicators are selected from a series of thermal comfort pointers, and the controllable indoor environmental parameters are analyzed and simulated on MATLAB to obtain the impact on the thermal comfort indicators, which can serve as important data to set up the fuzzy rule base. Next, according to the ISO7730 comfort standard and energy saving, three ways to control thermal comfort are proposed. With Arduino UNO as the development substrate, the sensing nodes for the indoor environment are set up, and the wireless sensing network is configured with ESP8266 to transmit the sensing data to the terminal. Monitored by the C# human-machine interface, the controllable load is controlled by wireless remote mode. Finally, the data is stored in the database for follow-up experimentation and analysis. Through actual measurement experiments, the thermal comfort and energy saving effects, under comfort, general, and energy-saving modes, as proposed in this study, are verified to achieve a balance between thermal comfort and energy saving.

Detecting Gear Surface Defects Using Background-Weakening Method and Convolutional Neural Network

Abstract: A novel, efficient, and accurate method to detect gear defects under a complex background during industrial gear production is proposed in this study. Firstly, we first analyzed image filtering and smoothing techniques, which we used as a basis to develop a complex background-weakening algorithm for detecting the microdefects of gears. Subsequently, we discussed the types and characteristics of gear manufacturing defects. Under the complex background of image acquisition, a new model S-YOLO is proposed for online detection of gear defects, and it was validated on our experimental platform for online gear defect detection under a complex background. Results show that S-YOLO has better recognition of microdefects under a complex background than the YOLOv3 target recognition network. The proposed algorithm has good robustness as well. Code and data have been made available.

Heterogeneous Gray-Temperature Fusion-Based Deep Learning Architecture for Far Infrared Small Target Detection

Abstract: This paper proposes the end-to-end detection of a deep network for far infrared small target detection. The problem of detecting small targets has been a subject of research for decades and has been applied mainly in the field of surveillance. Traditional methods focus on filter design for each environment, and several steps are needed to obtain the final detection result. Most of them work well in a given environment but are vulnerable to severe clutter or environmental changes. This paper proposes a novel deep learning-based far infrared small target detection method and a heterogeneous data fusion method to solve the lack of semantic information due to the small target size. Heterogeneous data consists of radiometric temperature data (14-bit) and gray scale data (8-bit), which includes the physical meaning of the target, and compares the effects of the normalization method to fuse heterogeneous data. Experiments were conducted using an infrared small target dataset built directly on the cloud backgrounds. The experimental results showed that there is a significant difference in performance according to the various fusion methods and normalization methods, and the proposed detector showed approximately 20% improvement in average precision (AP) compared to the baseline constant false alarm rate (CFAR) detector.

Shortest Path Routing Protocol Based on the Vertical Angle for Underwater Acoustic Networks

Abstract: Underwater Acoustic Networks (UANs) use acoustic communication. UANs are characterized by narrow bandwidth, long delay, limited energy, high bit error rate, and dynamic network topology. Therefore, UANs call for energy-efficient and latency-minimized routing protocol. In this paper, the shortest path routing protocol based on the vertical angle (SPRVA) is proposed. In SPRVA, the forwarding node determines the best next-hop according to main priority. When the main priorities of candidate nodes are the same, the alternative priority is used. The main priority is denoted by the residual energy and angle between propagation direction and depth direction, and the alternative priority is indicated by the link quality. SPRVA selects the node along the depth direction with more residual energy and better link quality as the best next-hop. In addition, a recovery algorithm is designed to avoid nodes in void areas as forwarding nodes. Simulation results show that SPRVA improves energy efficiency and decreases end-to-end communication delay.

A Novel Convolutional Neural Network Architecture for SAR Target Recognition

Abstract: Among many improved convolutional neural network (CNN) architectures in the optical image classification, only a few were applied in synthetic aperture radar (SAR) automatic target recognition (ATR). One main reason is that direct transfer of these advanced architectures for the optical images to the SAR images easily yields overfitting due to its limited data set and less features relative to the optical images. Thus, based on the characteristics of the SAR image, we proposed a novel deep convolutional neural network architecture named umbrella. Its framework consists of two alternate CNN-layer blocks. One block is a fusion of six 3-layer paths, which is used to extract diverse level features from different convolution layers. The other block is composed of convolution layers and pooling layers are mainly utilized to reduce dimensions and extract hierarchical feature information. The combination of the two blocks could extract rich features from different spatial scale and simultaneously alleviate overfitting. The performance of the umbrella model was validated by the Moving and Stationary Target Acquisition and Recognition (MSTAR) benchmark data set. This architecture could achieve higher than 99% accuracy for the classification of 10-class targets and higher than 96% accuracy for the classification of 8 variants of the T72 tank, even in the case of diverse positions located by targets. The accuracy of our umbrella is superior to the current networks applied in the classification of MSTAR. The result shows that the umbrella architecture possesses a very robust generalization capability and will be potential for SAR-ART.

A Method Combining CNN and ELM for Feature Extraction and Classification of SAR Image

Abstract: Due to the large number of Sigmoid activation function derivation in the traditional convolution neural network (CNN), it is difficult to solve the question of the low efficiency of extracting the feature of Synthetic Aperture Radar (SAR) images. The Sigmoid activation function in the CNN is improved to be a rectified linear unit (ReLU) activation function, and the classifier is modified by the Extreme Learning Machine (ELM). Finally, in this CNN model, the improved CNN works as the feature extractor and ELM performs as a recognizer. A SAR image recognition algorithm based on the CNN-ELM algorithm is proposed by combining the CNN and the ELM algorithm. The experiment is conducted on the Moving and Stationary Target Acquisition and Recognition (MSTAR) database which contains 10 kinds of target images. The experiment result shows that the algorithm can realize the sparsity of the network, alleviate the overfitting problem, and speed up the convergence speed of the network. It is worth mentioning that the running time of this experiment is very short. Compared with other experiment on the same database, it indicates that this experiment has generated a higher recognition rate. The accuracy of the SAR image recognition is 100%.

Surface Acoustic Wave Sensors for Ammonia Detection at Room Temperature Based on SnO2/Co3O4 Bilayers

Abstract: The selectivity of a SAW (surface acoustic wave) sensor, with a Co3O4 sensitive thin film for NH3 (ammonia) and the influence of SnO2 on its sensitivity, was studied. Thin films were deposited by pulsed laser deposition (PLD) on quartz SAW sensor substrates. Two sensors with different types of sensitive films were developed: a Co3O4 thin film sensor (S1) and a SnO2/Co3O4 thin film sensor (S2). The sensitive films were deposited in conditions which ensured a porous structure. The sensors were tested in the presence of three gases: NH3, methanol, and toluene. The selectivity of Co3O4 for NH3 was determined from the difference in the frequency shifts of the sensor for NH3 and for VOCs (volatile organic compounds). The positive influence of SnO2 on the sensitivity of sensor S2 was observed from the lower limit of detection (LOD) of this sensor and from the differences in frequency shifts between sensor S1 and sensor S2.

A Real-Valued Genetic Algorithm for Optimization of Sensor Placement for Guided Wave-Based Structural Health Monitoring

Abstract: The paper presents a novel implementation of the genetic algorithm (GA) to improve the coverage of the sensor network for damage detection using guided waves. The implementation allows depiction of sensor locations with real values which is closer to the real-life situation. Also, additional features such as proximity checks and node insertions have been implemented in order to improve the convergence of the GA as well as the thoroughness of the search space. For the traditional integer-based implementation, the size of the problem is large but finite. For the real-valued implementation, the problem size can indeed be infinitely large. So added measures have been introduced such as a two-step optimization process for the reduction in size and improved convergence.

Research on FBG-Based Sensor Networks and Their Coexistence with Fiber Optical Transmission Systems

Abstract: Market forecasts and trends for the usage of fiber optical sensors confirm that demand for them will continue to increase in the near future. This article focuses on the research of fiber Bragg grating (FBG) sensor network, their applications in IoT and structural health monitoring (SHM), and especially their coexistence with existing fiber optical communication system infrastructure. Firstly, the spectrum of available commercial optical FBG temperature sensor was experimentally measured and amplitude-frequency response data was acquired to further develop the simulation model in the environment of RSoft OptSim software. The simulation model included optical sensor network, which is combined with 8-channel intensity-modulated wavelength division multiplexed (WDM) fiber optical data transmission system, where one shared 20 km long ITU-TG.652 single-mode optical fiber was used for transmission of both sensor and data signals. Secondly, research on a minimal allowable channel spacing between sensors’ channels was investigated by using MathWorks MATLAB software, and a new effective and more precise determination algorithm of the exact center of the sensor signal’s peak was proposed. Finally, we experimentally show successfully operating coexistence concept of the spectrum-sliced fiber optical transmission system with embedded scalable FBG sensor network over one shared optical fiber, where the whole system is feed by only one broadband light source.