Visual detection for densely stacked small object (DSSO) has a wide range of applications in the construction, logistics, and import/export industries. Take the construction industry as an example, intelligent rebar counting, can considerably improve the management efficiency in sales, delivery and inventory management. It can also effectively prevent acts such as supervisory theft. However, current mainstream object detection models are susceptible to complex backgrounds when applying directly to DSSO detection, and have not been optimized for highly similar objects with high regularity of arrangement, resulting in low detection accuracy and efficiency. Therefore, we propose the DSSO-YOLO model, which introduces the coordinate attention mechanism into the C3 module of YOLOv5s to increase the accuracy of detection. And the bounding box loss function CIoU in YOLOv5s is replaced with Focal-EIoU-Loss, in order to accelerate the convergence speed and further improve the regression accuracy. To verify the performance of the algorithm, we constructed the steel products dataset SPDC. According to the test results conducted on SPDC, compared with YOLOv5s, the F1- Score, recall, mAP@0.5 and mAP@0.5:0.95 increased by 2.0%, 2.6%, 2.1% and 1.5%, respectively.

Few-shot learning is intended to address situations where there are few training samples per class. The main challenge is how to adequately extend the data or identify tighter associations from the sparse data. In this study, the final classification and data from few-shot learning are taken into account. First, we provide an efficient regularization to internal network generalization to accomplish data augmentation. Second, to enhance the Nearest Class Mean classifier (NCM) and soft K-means methods for classification, we raise a new regularized estimator based on the concept of Mahalanobis distance. We conduct extensive experiments based on CIFAR-FS and FC100. The ablation experiments show that the proposed data augmentation method is 2% higher than the baseline on CIFAR-FS(1-shot), and the classification algorithm is equally effective. The comparison experiments reveal results that much exceed the vast majority of state-of-the-art performance measures.

In the Three Gorges reservoir area, the overhead upright pier is the primary structural form. For intelligent monitoring of existing terminals, this research chooses Chongqing Xintian Port as the study object and proposes a support vector machine (SVM) damage-inducing factor (DIF) inversion model based on particle swarm optimization (PSO). To apply the finite element method to analyze the stress distribution characteristics of quay pile groups under three main DIFs, including the stacking effect, ship impact load effect, and bank slope effect. After characterizing the stress data, it becomes evident that there exists a correlation between stress and each DIF parameter. Before generating the training sample set, principal component analysis is employed to reduce dimensionality and eliminate a substantial amount of redundant data. The model has an accuracy of 0.999 for the identification of the type of DIF and 0.975 for the identification of the location of the action of the DIF with F1 coefficients of 0.999 and 0.978, respectively. For the strength of DIF predictions, MAE and MSE were 4.871 and 1.202, respectively, R2 was 0.986, NSE was 0.986, WI was 0.996, and PBIAS was 0.095. After extracting every sample, the relative error for the ship impact load effect is 0.05, and the highest relative error for the bank slope effect is 0.02; the error for the stacking effect is limited to 0.08. The results suggest that the damage inducement inversion model of the SVM optimized by the PSO algorithm can effectively identify the DIF of the overhead upright pier.

Uneven stowage of wharf is one of the main external loads during its service period. The main load-bearing structure of the overhead vertical wharf is the lower pile foundation. Under the action of the upper pile load, the pile foundation will be damaged to varying degrees, and the local damage may cause adverse effects on the overall structural safety of the wharf. In order to realize the inversion analysis of the adverse damage inducement of wharf stowage, the stress detection results of the lower pile foundation are taken as inversion data samples, and 10,000 sets of stress data are obtained by establishing a parameterized numerical calculation model. After normalization and dimension reduction, they are input into the established neural network model, and the action position and strength of stowage damage inducement are identified. The results show that the inversion analysis samples obtained by the parameterized model in this paper have higher accuracy and generalization in the calculation of neural network model.

Insertion/deletion polymorphisms (InDels) have been treated as a prospective and helpful genetic marker in the fields of forensic human identification, anthropology and population genetics for the past few years. In this study, we developed a six-dye multiplex typing system consisting of 34 autosomal InDels and Amelogenin for forensic application. The contained InDels were specifically selected for Chinese population with the MAF ≥ 0.25 in East Asia, which do not overlap with the markers of Investigator® DIPplex kit. The typing system was named as GoldeneyeTM DNA ID System 35InDel Kit, and a series of developmental validation studies including repeatability/reproducibility, concordance, accuracy, sensitivity, stability, species specificity and population genetics were conducted on this kit. We confirmed that the 35InDel kit is precise, sensitive, species specific and robust for forensic practice. Moreover, the 35InDel kit is capable of typing DNA extracted from forensic routine case-type samples as well as degraded samples and mixture samples. All markers are proved to be highly polymorphic with an average observed heterozygosity (He) of 0.4582. The combined power of discrimination (CPD) is 0.999 999 999 999 978 and the combined power of exclusion in duos (CPED) and trios (CPET) are 0.978 837 and 0.999573, respectively, which are higher than those of the Investigator® DIPplex kit. Thus, the GoldeneyeTM DNA ID System 35InDel kit is suitable for forensic human identification and could serve as a supplementary typing system for paternity testing. Supplemental data for this article is available online at https://doi.org/10.1080/20961790.2021.1945723 .

This paper describes the development and validation of a novel 31‐locus, six‐dye STR multiplex system, which is designed to meet the needs of the rapidly growing Chinese forensic database. This new assay combines 20 extended‐CODIS core loci (D3S1358, D5S818, TPOX, CSF1PO, TH01, vWA, D7S820, D21S11, D8S1179, D18S51, D16S539, D13S317, FGA, D1S1656, D2S441, D2S1338, D10S1248, D12S391, D19S433, and D22S1045), nine highly polymorphic loci in Chinese Han population (D3S3045, D6S1043, D6S477, D8S1132, D10S1435, D15S659, D19S253, Penta D, and Penta E), and two gender determining markers, amelogenin and Y‐Indel, which could amplify DNA from extracts, as well as direct amplification from substrates. To demonstrate the suitability for forensic applications, this system was validated by precision and accuracy evaluation, concordance tests, case sample tests, sensitivity, species specificity, stability, stutter calculation, and DNA mixtures, according to the guidelines described by the Scientific Working Group on DNA Analysis Methods (SWGDAM) and regulations published by the China Ministry of Public Security. The validation results indicate the robustness and reliability of this new system, and it could be a potentially helpful tool for human identification and paternity testing in the Chinese population, as well as facilitating global forensic DNA data sharing.

This paper proposes a damage identification method for piles by using two symmetrically arranged strain sensor arrays. Theoretical analysis shows that the curvature change have a significant change in the damaged region and can be used for damage localization, the relative curvature change in the damaged region is a constant and can be used to quantitative the severity of damage. The physical model experiments showed that the predetermined damage was located within the measured damage range, and the cross section damage coefficient had a linear correlation to the measured damage severity. The FEM results showed that the maximum relative error of damage severity is 3.3%, and the maximum error of damage location is 1 mm, the damage localization accuracy could be improved by increasing the strain measurement point density. The method is simple, does not rely on optimization algorithms, and is suitable for pile damage detection and monitoring.

Short tandem repeat within the male-specific part of the human Y chromosome (Y-STR) is an effective forensic tool in mixture identification, patrilineal relationship evaluation, and familial searches. Despite their usefulness, current Y-STR-based genotyping systems often lack the discriminatory power to resolve genetic relationships between distant relatives or within patrilocal populations. In this study, we developed a novel Y-STR 29-plex typing system, which combined the 17 Y-STR loci used in the AmpFLSTR® Yfiler® PCR Amplification Kit (Yfiler), eight Y-STR loci with a low-medium mutation rate, and four rapidly mutating Y-STR loci. The system was generated to achieve greater discriminatory power between male subjects and improved ability to infer haplogroup classifications. The system was extensively tested on data from 752 individuals for its sensitivity, male specificity, species specificity, mixture resolution, reproducibility, concordance, stutter and size accuracy, precision, and population genetics, following the Scientific Working Group on DNA Analysis Methods (SWGDAM) guidelines. The results demonstrated that the Y-STR 29-plex typing system was time-efficient, reproducible, accurate, sensitive, and robust to familial searching and paternal biogeographic ancestry inference.

Memristors have emerged as a potential tool to implement the training and operation of an integrated neural network, because of its current-voltage curve of the hysteresis loop and unique pulse regulation resistance method. However, most of the existing neural networks implemented on memristors are relatively basic architecture, and the processing functions are limited to the recognition of the simple signal and image models. In this paper, we propose a 3D Convolutional Neural Network based on memristor to recognize and classify the behaviors of human in the video with 6 main actions. As an extension of 2D Convolutional Neural Networks, 3D Convolutional Neural Networks have attracted attention for video information processing, since it introduces the time dimension innovatively on the basis of spatial dimensions to capture the contextual information between the different frames in the video. Accordingly, we use the 3D Convolution to construct our proposed neural network based on memristors. Besides, we use the basic 3 x 3 memristor arrays to construct the larger functional memristor arrays and form the 3D convolutional layers of our network by considering that the 3 x 3 basic memristor array has excellent flexibility and anti-jamming capability. With this strategy, we can make full use of the hardware structure to improve accuracy while reducing hardware noise. Finally, we implemented network obtain more than 70% accuracy on the Weizmann video dataset. This demonstration is an important step that memristors can implement the much larger and more complex neural networks for processing the more complex applications. (C) 2018 Elsevier B.V. All rights reserved.

To ensure the safety of the navigation, self-propelled ship model test is widely used in navigable administer engineering to visually and really reflect navigable condition, offering reasonable suggestions for the design of route. This paper proposed a video analysis based route tracking approach for self-propelled ship model that sails in large scale river models, which can realize automatic measurement for ship model motion parameters. Firstly, the captured videos of the self-propelled ship model are transferred to the computer via wireless local area network (WLAN). Then, the camera lens distortion is eliminated by rectify and aerial view is reconstructed. Third, ORB and binary BoF classifier are used to detect ship model. At last, through frame difference and Freeman chain-code, the coordinates of ship model's markers can be obtained. Moreover, on visual interactive interface, the route of the ship model is plotted, and velocities and drift angles of ship model are also calculated. This approach was tested on the several river models, such as Jianzishan navigation junction, which is a medium-sized water conservancy project located in the middle reaches of the Minjiang River. The results from the experiments demonstrate the approach can not only realize accurate measurement of coordinates of ship model, but also can visually map out the route of the ship model that sailing in large scale river models.

The performance degradation of fiber Bragg grating strain sensors is unavoidable under stress fatigue,in order to assess the ultimate sensing life of fiber Bragg grating sensors,the full scale relative error expression of stress fatigue effect is deduced from the principle of fiber Bragg grating,and the industrial instrument accuracy class is used as the limit state threshold of the sensor.And then a method based on the beam of constant strength is put forward to test the ultimate sensing life.The ultimate sensing life of fiber Bragg grating in high pile wharf structural health monitoring application is assessed by this method,the failure threshold of full scale relative error is set as 4%,4 fiber Bragg grating sensors suffer from 132 million times alternate strain,when the fatigue number reach to 132 million times,2 sensors reach the ultimate sensing life,1 sensor proximity reach the limit.That is,the maximum sensing life is 132 million times.Experimental results show that the proposed method can effectively test the ultimate sensing life of stress fatigued fiber Bragg grating strain sensors in different structural health monitoring application.

针对表面粘贴式非栅区封装结构,依据变形等效原理,推导了该结构的应变传递函数,仿真分析了非栅区封装中间段长度与封装材料弹性模量对应变传递效率的影响.分别采用有机环氧胶与金属锌,将两只光纤布喇格光栅用非栅区封装方式固定在同一根钢丝上,进行拉伸试验,试验结果表明:两种材料的拉伸曲线线性度均达到0.99以上,表明两种封装材料的应变传递一致性较好;金属锌封装结构的应力感知灵敏度平均值为0.142 6 nm/KN,环氧有机胶封装结构的应力感知灵敏度平均值为0.130 4 nm/KN;各种应变值下,金属锌封装结构的应变传递系数均稳定在0.995左右,而环氧有机胶封装结构的应变传递系数在0.91左右上下波动,金属锌的应变传递效率较有机环氧胶高了近9.34％.试验结果与数值仿真结果较为吻合,可为非栅区封装光纤布喇格光栅传感器的设计提供依据.

弹性介质在受到外力的情况下会发生弹性形变.在弹性介质质量一定的情况下,此弹性形变会使弹性介质的体积和密度伴随着改变.对于超声波而言,在除了弹性介质本身的其他因素不变的情况下,弹性介质体积和密度的改变将会影响到超声波的传播特性.通过理论分析得出超声波幅值和相位随弹性琼脂应力变化的关系.以琼脂作为实验模型,检测琼脂在形变时超声波的信号参数,并且通过实验数据,最终得到二者的曲线和标定公式.此结果有利于今后在介质应力检测及超声波检测技术等方面的深入研究.

Bonding layers, serving as the strain transmission mediums, may bring undesirable effects to the sensing properties of fiber Bragg grating (FBG) strain sensors during their fatigue process. To analyze the strain sensitivity and the reflected spectrum of FBG strain sensors in different fatigue stages of bonding layers, their strain sensitivities were derived according to strain transfer models. Resorting to T-matrix formalism, the affected reflected spectra were stimulated. Theoretical analysis results show that there is a gradual decline in the strain sensitivity during the stage of fatigue crack initiation, and that significant distortions emerge in the reflected spectrum during the stage of fatigue crack propagation. In addition, a cyclic loading fatigue test was conducted and the phenomenon observed in the test showed a good agreement with the theoretical prediction. (C) 2014 Society of Photo-Optical Instrumentation Engineers (SPIE)

Cable anchor is one of the most important parts of cable-stayed bridge. Stress distribution in cable anchor reflects the status of the bridge, but it is very difficult to measure. Embedded fiber Bragg grating sensors provide a good solution. This paper first describes the structure of cable anchor in detail and analyzes the stress distribution of cable anchor in theory. And then, a scheme based on carbon fiber reinforced embedded FBG sensors is proposed. Finally, an experiment on full-scale cable anchor is conducted to evaluate the feasibility of the proposed non-destructive method. The experimental results show that the non-destructive testing method based on embedded FBG sensors can measure the stress distribution in cable anchor effectively.

[This corrects the article DOI: 10.1371/journal.pone.0057471.].

Metal bonding layer seriously affects the strain transfer performance of Fiber Bragg Grating (FBG). Based on the mode of FBG strain transfer, the influence of the length, the thickness, Poisson’s ratio, elasticity modulus of metal bonding layer on the strain transfer coefficient of FBG is analyzed by numerical simulation. FBG is packaged to steel wire using metal bonding technology of FBG. The tensile tests of different bonding lengths and elasticity modulus are carried out. The result shows the strain transfer coefficient of FBGs are 0.9848,0.962 and their average strain sensitivities are 1.076 pm/με,1.099 pm/με when the metal bonding layer is zinc, whose lengths are 15mm, 20mm, respectively. The strain transfer coefficient of FBG packaged by metal bonding layer raises 8.9 percent compared to epoxy glue package. The preliminary experimental results show that the strain transfer coefficient increases with the length of metal bonding layer, decreases with the thickness of metal bonding layer and the influence of Poisson’s ratio can be ignored. The experiment result is general agreement with the analysis and provides guidance for metal package of FBG.

Steel cable plays an important role in modern infrastructure due to its special characteristics. Because most of structure load is transformed to the cable tension in cable stayed structures, it is very important to monitor cable tension. Being a slender element, Fiber Grating Strain Sensor is sensitive to axial strain and is regarded as a most prospective way to monitor the cable tension. The paper reviews a series of problems of FBG when embedded into the cable, and introduced five different embedded FBG strain sensors. Principle, characteristics, and application states of these five sensor has been discussed in details. The prospective of embedded FBG Strain Sensor for cable tension has been forecast.

To simplify strain transfer links and improve long-term strain transfer stability, a novel strain measuring method based on metallized bonding is proposed. The strain sensing structure is proposed, and the basic principle and technical process of metallized bonding technology are analyzed. According to the elastic mechanics control equation, the stress transfer model for the optimal selection of metallized bonding materials is established. Based on the stress transfer model, the influence of different bonding materials on strain sensing performance is analyzed in simulation, the simulation results prove the theoretical superiority of direct metallized bonding. To test the sensing property of direct metallized bonding, metal Pb was selected as the bonding material, a FBG was bonded directly onto a steel string with diameter of φ 7 mm, and tensile tests were carried on a material test system (MTS). The experimental results show that good linearity and repeatability exist between fiber Bragg grating (FBG) output central wavelength and tensile force, the linearity is better than 0.999; and the average strain transfer coefficient reaches 0.98, it is in good accordance with the theoretical value of 0.9828, which proves the correctness and feasibility of the theoretical model and bonding technical approach. The test results indicate that the FBG strain sensing method based on direct metallized bonding has superior strain measurement capability.

To improve the performance of packaged FBG strain sensors, bond mode between FBG and its protecting shell was optimized. Based on the simplified strain transfer model, the strain transfer efficiency K of bond layer is deduced. Theoretical simulation reveals that strain transfer efficiency and stability of bonding layer are strongly dependent on the Young's modulus of bond layer, and it is better to replace organic adhesives with metal as the bonding layer. To address the problem of metal bonding, a special metallized bonding device is developed according to sputtering. To verify the validity of theory and methodology, Pb is selected as the bonding material, and a FBG was bonded directly onto a steel rod with a diameter of 7 mm, and bonding effect was tested on a Material Test System (MTS). The experimental results showed that there was no delamination in bonding layer when strain increasing up to 3500 mu epsilon. Experimental results also revealed a strain transfer efficiency of 0.98 with linearity of 0.999, which was very well accordance with theoretical value of 0.9828.