Search

Cracks are the main form of tunnel complications and have drawn much attention in current tunnel safety monitoring. However, automatic crack detection cannot accurately extract cracks from tunnel concrete lining surfaces because of the inference of noises like lining seams, which have similar gray values and textures to cracks, and increase the topological complexity of cracks. In this paper, we proposed a novel method for eliminating lining seams in tunnel concrete crack images, to address the above-mentioned issues. Our contributions are shown as follows: 1) classified arbitrary line segments to better distinguish the edges on the cracks and the seams, and proposed uniform k-divided angle model to get accurate classification and 2) introduced two new principles to help effectively eliminate lining seams through the information of translation direction and expansion width of all line segments. The experimental results proved the superior precision and efficiency of our method compared with the existing methods.

To make up for the deficiencies in the present method of measuring microdetonator output energy, this paper uses PVDF (Polyvinylidene Fluoride) piezoelectric films as sensors and designs a system to test the output pressure of microdetonators with a 2.5 mm diameter. The average measured output pressure was 9.87GPa. In addition, detonation tests were carried out which showed that a microdetonator with a 10 mg charge could reliably detonate a JO-9C(III model) lead explosive with an 8 mg charge in a 0.7 mm air-gap. The detonation simulations for lead explosives using LS-DYNA software were consistent with the detonation test results. There was less than 5% error between the output pressure values and the test values of microdetonators. The results showed that PVDF piezoelectric films have rapid response times, high precision and low production cost in output pressure tests. The simulation model has practical significance for the quantitative evaluation of microdetonator output energy.

The manufacturing process of planar microsprings by the UV-LIGA process was introduced; the mechanism of fabrication error was analyzed from three process stages of exposure, SU-8 swelling, and corrosion; and the corresponding improvement measures were proposed. The line width of two groups of microsprings after the exposure, SU-8 swelling, and debonding was measured, and the elastic coefficients in three coordinate directions were accurately measured. The results show that the factors that affect the line width error are stable. The effects of the exposure diffraction, swelling of SU-8 glue, and debonding corrosion process on the line width size follow the order: corrosion error > swelling error > exposure error. When the line width is small, the aspect ratio of the microspring is large, and the direction of machining error is generally negative. The relative error of the elastic coefficient of microsprings in three coordinate directions is larger. The influence rule of the line width error on the elastic coefficient follows the order: kerrory > kerrorx > kerrorz.

In this paper, the classic template matching sequential similarity detection algorithm(SSDA) has a large amount of calculation and slow operation time, and an improved SSDA is proposed. By introducing the integral image, the point-by-point summation in the matching template window is transformed into the addition and subtraction of four positions in the integral image, which reduces the amount of calculation for point-by-point accumulation and summation. The laser image and the classic lena image are used as the experimental images to be matched, and the image with gray scale changes and noise added as the template image is used for image matching. Compared with the traditional normalized gray cross-correlation(NCC) algorithm and the improved NCC algorithm for integrating images, the experimental results show that the improved algorithm can shorten the running time, have a certain anti-interference ability against noise, and achieve image matching.

With the traditional interactive method of virtual scene, when users observe art design works of virtual scene environment, they usually cannot understand where they are at the moment and feel lost. Therefore, a new interactive method of virtual scene in environment art design is proposed. The design model of the environment scene is introduced. The design process of the interactive module is introduced, and the location of the user in the three-dimensional scene is determined. The realization process of the interactive communication module is analyzed, and the implementation mechanism and process of virtual scene interaction are introduced. In the two-dimensional map, the plane coordinates are transmitted to the virtual scene, then using the relocation function to get the coordinates in the virtual scene, and display virtual scenes centered on this coordinates to achieve the purpose of interactive virtual scene. Experimental results show that the proposed method can effectively realize virtual scene interaction without loss of memory, and the hardware consumption and real-time testing can meet the requirements of user vision and interactive experience, strong robustness, high degree of user interest.

Trusted execution environment (TEE) is a promising technique to protect user programs and data on public cloud environments. To support unmodified applications running, many TEE runtime systems have been proposed. However, a major drawback of the existing schemes is the lack of interface protection. This problem may lead to many security problems, such as memory information leakage and malicious codes attacks. To tackle this problem, we propose SeRoT, a new secure runtime system on trusted execution environments. Our secure runtime system first provides some core functions to the enclave programs. Then we protect the host interface at two levels, binary interface level and application interface level. In these two levels, we prevent the adversary interfacing with malicious messages. Furthermore, we implement SeRoT on a RISC-V based platform and show our scheme is average about 10% faster than Keystone on two popular and representative benchmarks.

With the rapid development of information technology, the talents of traditional industries can not adapt to the development of social economy very well. How to innovate the training ideas and reform the educational and teaching plan is a problem that needs serious thinking in the era of information technology. Combined with the strategic guidance of the state on the application-oriented transformation and development of universities in recent years, this paper focuses on the training of architectural talents, and analyzes the present situation of architectural talents training. From the perspective of innovation-driven development, the training goal and training ideas of applied talents are put forward.

This report describes a novel therapeutic technique called photo-mediated ultrasound therapy (PUT). PUT applies synchronized short pulse duration (nanosecond) laser and ultrasound burst on targeted tissue, offering high-precision localized treatment. PUT is based on controlled induction and promotion of micro-cavitation activity in the target tissue. PUT is able to safely and effectively treat retinal neovascularization in rabbits with persistent nonperfusion up to 4 weeks after PUT in the choroidal vasculature.Clinical Relevance- PUT can selectively remove retinal angiogenesis in clinically-relevant disease models in humansized eyes (rabbit) without damaging surrounding tissue.

Harmonic and subharmonic components are contained in the spectrum of the steady-state visual evoked potential (SSVEP). These components are widely used in the brain-computer interface (BCI) technology and neuroscience research. Subharmonics cannot be easily induced, and the mechanism of the power variation of these components are ambiguous. In this paper, the induced factors and power variation of the SSVEP subharmonics are studied by a visual cortex network model simulation. The SSVEP subharmonics with orders higher than 1/2 are easily suppressed by noise. Therefore, the power variation of the SSVEP subharmonics with 1/2 order, and their induction factors are further analyzed. Simulation results show that the stimulus frequency, stimulus intensity, and noise intensity are three important parameters affecting the generation and power variations of the SSVEP subharmonics. If the stimulus intensity is weak, the 1/2 subharmonics cannot be induced. When the stimulus intensity increases to a certain value, 1/2 subharmonics appear in the SSVEP, and their power increases sharply with the increasing stimulus intensity. However, when the stimulus intensity exceeds a certain threshold, the 1/2 subharmonics remain constant. The bandwidth of the stimulus frequency, which can induce 1/2 subharmonics increases with increasing stimulus intensity under the same noise condition. Generation of 1/2 subharmonics and its power are suppressed by noise.

Imbalanced data exists commonly in machine learning classification applications. Popular classification algorithms are based on the assumption that data in different classes are roughly equally distributed; however, extremely skewed data, with instances from one class taking up most of the dataset, is not exceptional in practice. Thus, performance of algorithm often degrades significantly when encountering skewed data. Mitigating the problem caused by imbalanced data has been an open challenge for years, and previous researches mostly have proposed solutions from the perspectives of data re-sampling and algorithm improvement. In this paper, focusing on two-class imbalanced data, we have proposed a novel divergence-encouraging autoencoder (DEA) to explicitly learn features from both of the two classes and have designed an imbalanced data classification algorithm based on the proposed autoencoder. By encouraging maximization of divergence loss between different classes in the bottleneck layer, the proposed DEA can learn features for both majority and minority classes simultaneously. The training procedure of the proposed autoencoder is to alternately optimize reconstruction and divergence losses. After obtaining the features, we directly compute the cosine distances between the training and testing features and compare the median of distances between classes to perform classification. Experimental results illustrate that our algorithm outperform ordinary and loss-sensitive CNN models both in terms of performance evaluation metrics and convergence properties. To the best of our knowledge, this is the first paper proposed to solve the imbalanced data classification problem from the perspective of explicitly learning representations of different classes simultaneously. In addition, designing of the proposed DEA is also an innovative work, which could improve the performance of imbalanced data classification without data re-sampling and benefit future researches in the field.

Voltage regulation depending on reactive power compensation is the main feature of the AC power supply system. Magnetically controllable reactors (MCR) are becoming a growing demand for this purpose. The structure and working principles of MCR are analysed in this paper while a simulation model is established. The reactive power compensation strategy based on a single loop voltage control system (SLVCS) is presented and a double loop voltage-current control system (DLVCCS) is proposed. A comprehensive scenario is developed to mitigate reactive power compensation by using the proposed controls. Simulation results substantiate that proposed controls of MCR has a faster response in comparison to the traditional control, and it provides the least voltage variation at the line end. It also shows that the proposed control on MCR meet the desired objective of the voltage regulation and provides flexibility to ac transmission system.

The construction of telecommunication network is a complex engineering, requiring telecom operators to focus on the operational effects after construction, as well as summarize the construction experiences for further improvements. This paper innovatively proposes indicators framework of network construction and develops a post-evaluation model based on analytic hierarchy process (AHP) method. Furthermore, the proposed post-evaluation model is used to analyze the network operation effect after construction in two cities, the result verifies that the proposed post-evaluation model has promising performance and practical effect for operators.

The 7B04-T6 and 30CrMnSiA joints were pretensioned and exposed to hot and humid marine climate for corrosion test. They were taken back after different cycles of test and fatigue-loaded. Then the fatigue life values of six cycles were studied. The fatigue-safe life of the joints was estimated by using the one-sided tolerance factor method. The results show that the fatigue life of the joint specimen is greatly reduced by the interaction of the corrosive environment and fatigue load. Taking the confidence and reliability into consideration, the one-sided tolerance factor method makes it is possible to obtain an accurate estimated value of fatigue-safe life and thus plays a guiding role in the use of engineering application materials.

Research on performance trend and life assessment of glass fiber reinforced composite in storage condition. Develop four different temperature of thermal oxidative aging test on glass fiber reinforced composite, detect impact property of composite regularly. Use life assessment method to deal with the test data and evaluate the trend of material property. Through calculation, we can get storage life of composite in tropical ocean climate and dry heat desert climate environment is from 14.99 to 31.26 years. Test result shows the composite has good storage performance. Its life could more than 14.99 years in atrocious tropical ocean climate. And the test and evaluate method is suitable for assessing the composite life.

Internet and cloud computing are developing rapidly. Nowadays malicious network users often conceal attacks as ordinary access, but intrusion detection systems can not completely distinguish them. This paper proposed a decision model for network defenders of honeypot-enabled system. It firstly modeled the interaction between the malicious user and the defender as repeated games, and depicted the uncertainty behaviors of the malicious users by Bayesian model. Then based on the relative historical payoffs of game players, a Bayesian improved model is proposed. By the model defender decides whether to lead the visitor to regular services or to honeypots, which increases the attacker's cost and reduces attacks finally.

Neural networks with both high accuracy and small network size are urgently required for mobile phone applications. However, previous network search methods do not take network size into account. In this paper, we use the reinforcement learning method to search for networks offline, with both high accuracy and small network size. Gaussian policy is used to explore the number of convolutional channels in a finer manner. Parameter reward is included in our reward function to punish large networks. We also use binary networks to further reduce network size. Without skip connections or branches, the network generated by our method is competitive with other methods on Cifar-10. Our network is much smaller than networks generated by other network search methods. Besides, our accuracy is higher than original binary network reported in BinaryConnect and is competitive with other real-valued networks.

Standard convolutional neural networks assemble multiple convolutional layers to extract high-level features. Recent efforts keep designing deeper and wider architectures. Even with skip connections applied to combine different layers, the useful low-level features are not effectively utilized. Some deep layer aggregation methods have been proposed to aggregate features of all levels, using simple linear combination or complex non-linear transformation. In this paper, we treat convolutional features as a sequence, and propose our Recurrent Aggregation of Convolutional Neural Network (CNN-RA). Our aggregation method splits a standard CNN into blocks and maps their feature matrices to a sequence of vectors of the same length. LSTM is employed to connect to the sequence and better fuse features across blocks. Our proposed CNN-RA can be directly appended to any standard CNN without any modifications. Experiments show remarkable improvements of CNN-RA over the original architectures across datasets.

The performance of DC transmission system is determined by DC control strategy and control parameters. Appropriate control strategies and parameters can not only improve the operating characteristics of the DC system itself, but also improve the stability level of AC system. In this paper, the parameters of the voltage dependent current order limiter (VDCOL) which affect the stability of the AC / DC system are analyzed and the modified VDCOL parameters are proposed in view of the AC faults in the DC receiving end. The influence of emergency DC power support on the stability of AC / DC system is analyzed in view of the incoming DC fault in AC / DC system, and the sensitivity index of emergency DC power support is proposed. Emergency DC power support and VDCOL coordinated recovery control for AC / DC systems are proposed. A simulation system based on the actual large power grid in Shandong is used for validation. The simulation software is based on the PSD-BPA program.

Anomaly detection is one of the basic problems in the field of data mining. It has been concerned and widely used by industry and academia. With the continuous increase of hydrological data, the current anomaly detection algorithm is too low in time efficiency. Besides, there are too many anomaly points excavated. In the face of so many anomaly points, the analysis decision-makers have no way to start. For this problem, this paper uses the isolation forest algorithm for the hydrological data of the pattern representation, and a hydrological time series anomaly pattern detection algorithm based on isolation forest is proposed. At the same time, it is difficult to determine the partition threshold in isolation forest and can not output top-k. K-means clustering respectively algorithm and nearest neighbor algorithm are used to improve the isolation forest algorithm, which can effectively overcome the subjectivity of artificially setting threshold and improve the stability of results expression. It is applied to the measured data of the Chuhe River Basin, and compared with other improved algorithms in accuracy and time complexity. The effectiveness of the improved isolation forest algorithm is verified by experiments.