Table of Content

31 July 2019, Volume 37 Issue 4

Previous Issue    Next Issue

Communication Engineering

Transmission-Line Galloping Monitoring Based on Phase-Sensitive Optical Time-Domain Reflectometry

HAO Weibo, ZHAO Yanshuang, LI Zhuoshu, XIE Kai, XING Jian, ZHANG Jianzhong, YUAN Libo

2019, 37(4):  437-446.  doi:10.3969/j.issn.0255-8297.2019.04.001

Asbtract ( 972 )   PDF (7156KB) ( 36 )  

References | Related Articles | Metrics

The phase-sensitive optical time-domain reflectometer (φ-OTDR) is proposed to monitor the galloping of the optical phase conductor (OPPC). Optical fiber cables embedded inside OPPC is used as sensing elements. As the exciter of the galloping machine applies different excitation frequencies on the OPPC, the measurement signals of φ-OTDR system are analyzed to figure out galloping characteristics. For different excitation frequencies, the φ-OTDR system measurement signals have corresponding frequency characteristics.

Signal and Information Processing

Compound Control Strategy Based on LCL Filter

XUAN Zhaoyan, JIA Wanyong, CHEN Xuebin, JING Huicheng, ZHAO Xin, MA Zhenyu

2019, 37(4):  447-458.  doi:10.3969/j.issn.0255-8297.2019.04.002

Asbtract ( 855 )   PDF (3910KB) ( 65 )  

References | Related Articles | Metrics

LCL type grid-connected inverters, when directly adopting the proportionalintegral (PI) and repetitive control series or parallel control strategies, generally suffer the problem of controllers interfering with each other. To overcome the problem, we propose a series-parallel composite control strategy for grid-connected inverters in this paper. This strategy combines the dynamic response of PI control with strong anti-interference ability of quasi-proportional resonance (QPR) and high steady-state accuracy of repeating control. The PI controller performs fast error-tracking dynamic state, whereas in steady state, no static error compensation is required by using repeating control and QPR. The design principle of each controller is analyzed in detail, and the Matlab simulation model of each controller is established. Simulation results show that the series-parallel composite control scheme has better dynamic response speed and steady-state compensation performance.

Radial Basis Network Training Algorithm Based on Surface-Simplex Swarm Evolution

WEI Wei, QUAN Haiyan

2019, 37(4):  459-468.  doi:10.3969/j.issn.0255-8297.2019.04.003

Asbtract ( 1004 )   PDF (3066KB) ( 75 )  

References | Related Articles | Metrics

The control parameters of intelligent optimization algorithm have a great influence on the learning performance of intelligent optimization algorithm. In order to solve this problem, a radial neural network training algorithm based on simplex evolution is proposed. It uses a one-dimensional neighbor based full random search method to reduce the number of control parameters, maintain the particle diversity through the group multicolor state, and avoid the algorithm falling into the local extremum point. Simulation results show that the algorithm not only improves the recognition rate but also reduces the influence of control parameters on learning performance. The generalization and robustness of the algorithm are improved.

Tagged Visual Cryptography Scheme Based on XOR Decryption

XU Mengqi, Lü Donghui, REN Yanli

2019, 37(4):  469-480.  doi:10.3969/j.issn.0255-8297.2019.04.004

Asbtract ( 810 )   PDF (5776KB) ( 36 )  

References | Related Articles | Metrics

Tagged visual cryptography is a new type of visual cryptography, which can hide tags into each generated common share to provide users with supplementary information. The existing tagged visual cryptography has some problems such as poor visual quality, pixel expansion and design of encryption matrices. Therefore, this paper proposes an improved tagged visual cryptography based on XOR decryption. In encryption process, the scheme embeds tagged information in the generated common share; and in decryption process, XOR operation is used to recover the secret image. The scheme decrypts the image without pixel expansion problem, and the requirement of design of encryption matrices. At the same time, it can adjust the probability of embedded tagged information. Compared with the existing tagged visual cryptography, the image recovered by the scheme is clearer.

Anatomical Landmark Localization in Lateral Cephalograms by Using Two-Layer Regression Forests

QIN Zhen, DAI Xiubin, XIE Lizhe

2019, 37(4):  481-489.  doi:10.3969/j.issn.0255-8297.2019.04.005

Asbtract ( 733 )   PDF (4906KB) ( 28 )  

References | Related Articles | Metrics

To automatically detect anatomical landmarks in cephalometric X-Ray images, a context-aware landmark detection method using two-layer regression forest models is proposed. First, it extracts appearance features from images to train the first-layer regression forest model, which can be used to generate a displacement map for each landmark per training image. Second, from the displacement maps, the context features are computed and combined with appearance features to train the second-layer regression forest. Then, by exerting the trained two-layer regression forest model on the new cephalometric X-Ray images to be processed, the displacement vectors of all pixels to each target landmark will be produced. Finally, the proposed method uses regression voting to acquire the landmark position in the testing image. Experimental results show that the proposed method has good performance in the detection of cephalometric landmarks in dental X-Ray images.

Grade Determination of Fine Grain Ore Based on Depth Image Analysis

LU Caiwu, QI Fan, RUAN Shunling

2019, 37(4):  490-500.  doi:10.3969/j.issn.0255-8297.2019.04.006

Asbtract ( 1117 )   PDF (7482KB) ( 72 )  

References | Related Articles | Metrics

To improve the precision of image processing technology in fine-grained ore measurement, a grading determination method based on depth image analysis is proposed. On the basis of gray-level co-occurrence matrix (GLCM), a maximum linear dispersion method is proposed to generate the step size and gray-level compression level adaptively, and support vector machine (SVM) classifiers are optimized through grid search and crossvalidation thus to improve the accuracy of particle size measurement. Experimental results showed that this method can achieve an accuracy rate of more than 92% for fine-grained ores with particle sizes of 0~0.9 mm, 0.9~3.0 mm, 3.0~5.0 mm and 5.0~7.0 mm, which can fully meet the requirements of grading and determining fine-grained ores.

An Image Caption Generation Model Combining Global and Local Features

JIN Huazhong, LIU Xiaolong, HU Zike

2019, 37(4):  501-509.  doi:10.3969/j.issn.0255-8297.2019.04.007

Asbtract ( 1056 )   PDF (5770KB) ( 125 )  

References | Related Articles | Metrics

An image caption generation model with attention mechanism combined with local and global features is proposed for dealing with the weakness of the image description model by the local image features. Under the framework of encoder and decoder architecture, the local and global features of images are extracted by using Inception V3 and VGG16 network models at the encoder, and the image features of two different scales are fused to form the coding results. On the decoder side, long short-term memory(LSTM) network is used to translate the extracted image features into natural language. The proposed model is trained and tested on Microsoft COCO dataset. The experimental results show that the proposed method can extract more abundant and complete information from the image and generate more accurate sentences, compared with the image caption model based on local features.

Algorithm for Extracting and Tracking Rainstorm Events Based on Time Series Raster-Formatted Datasets

YANG Guanghui, XUE Cunjin, LIU Jingyi, WU Qunyong, WU Chengbin

2019, 37(4):  510-518.  doi:10.3969/j.issn.0255-8297.2019.04.008

Asbtract ( 708 )   PDF (5512KB) ( 73 )  

References | Related Articles | Metrics

Based on evolution characteristics of rainstorms, we propose an algorithm for extracting and tracking rainstorm events with long-term raster-formatted datasets. The first step is the rainstorm extraction in the temporal domain. We calculate the cumulative rainfall of each grid of a time series, and identify the rainstorm levels. The second step is the rainstorm connection in a spatial domain:connecting the adjacent spatial grids tagged by rainstorm in a rainstorm object, and transforming it from raster format to the vector one with temporal, spatial and thematic information. Then rainstorm is tracked in the spatiotemporal domain. The relationship between storm objects will be fined based on the intersection of spatial topologies at adjacent moments, and the rainstorm objects are extracted based on the relationship, and calculate rainstorm event information. We using GPM-IMAGE final products, station datasets and radar datasets to compare and verify the algorithm. Experimental results show that the algorithm can completely extract the dynamic development process of rainstorm.

Trend Analysis of Vegetation Cover Changes Based on Spearman Rank Correlation Coefficient

WANG Dianlai, SU Aixia, LIU Wenping

2019, 37(4):  519-528.  doi:10.3969/j.issn.0255-8297.2019.04.009

Asbtract ( 850 )   PDF (3557KB) ( 163 )  

References | Related Articles | Metrics

Spearman rank correlation coefficient method is proposed and its feasibility and applicability are also investigated, in view of problems that Pearson correlation coefficient method suffers noise sensitivity and limited finding ability of linear relationship in the longterm trend analysis of vegetation cover changes. Firstly, the anti-noise ability of Spearman rank correlation coefficient method is studied by simulation. Secondly, based on SPOT vegetation normalized vegetation index (NDVI) data from 1998 to 2013, Pearson correlation coefficient, Mann-Kendall test and Spearman rank correlation coefficient method are used to detect the vegetation cover changes in Inner Mongolia, and the results are graphically presented. The differences of the three methods are compared. The experimental results show that Spearman rank correlation coefficient has better anti-noise performance. There is high consistency in spatial distribution of vegetation cover changes among the results of Spearman rank correlation coefficient, Pearson correlation coefficient, and Mann-Kendall test. The maximum difference in vegetation increase and decrease regions does not exceed 2% in three methods.

Computer Science and Application

Urban Passenger Flow Aggregation Risk Forecasting Based on Principal Component Regression Algorithm

WANG Juquan, WANG Wei, MA Huimin, YANG Bo, DU Wen

2019, 37(4):  529-540.  doi:10.3969/j.issn.0255-8297.2019.04.010

Asbtract ( 714 )   PDF (2576KB) ( 59 )  

References | Related Articles | Metrics

In order to solve the problem of the low accuracy of early warning of public events in mega-cities, this paper proposes a principal component regression algorithm to fit the mobile user data and real passenger flow data of fixed regions provided by operators, and uses a variety of statistical methods to test and evaluate the model. The principal component analysis can effectively overcome the multicollinearity problem of the mobile phone user data provided by the operators. By making full use of all-dimension information of the mobile phone user data, the complexity of the algorithm is reduced, and the accuracy of the prediction of the urban passenger flow aggregation risk is effectively improved.

Convolutional Neural Networks Text Classification Model Based on Attention Mechanism

ZHAO Yunshan, DUAN Youxiang

2019, 37(4):  541-550.  doi:10.3969/j.issn.0255-8297.2019.04.011

Asbtract ( 1214 )   PDF (1946KB) ( 403 )  

References | Related Articles | Metrics

Text categorization is an important part of natural language processing. Effective extraction of global semantics is the key to the success of text categorization. In order to emphasize the non-local importance of the extracting feature of convolutional neural networks, an A-CNN text classification model including four Attention CNN layers is established by using Attention mechanism. In the A-CNN model, the general convolution of the Attention CNN layer is used to extract local features, and the Attention mechanism is used to generate feature non-local correlation. Finally, the A-CNN model is experimentally used for the analysis on data sets such as sentiment analysis, problem classification, and question answer selection. Compared with other models, the A-CNN model improves the classification precision of the three above tasks by 1.9%, 4.3%, and 0.6%, respectively. The A-CNN model performs higher accuracy in text classification tasks and stronger versatility.

Control and System

Formation Control for Multi-UAV Systems Based on Radom Time-Delay

JI Lei, FAN Chunxia

2019, 37(4):  551-564.  doi:10.3969/j.issn.0255-8297.2019.04.012

Asbtract ( 618 )   PDF (3164KB) ( 90 )  

References | Related Articles | Metrics

During the flight of formatted multi-unmanned aerial vehicles, interactive communication among the individual is necessary over a communication network. In the communication, transmission time-delays may occur due to the network limits on communication bandwidth, the amount of traffic data and others. In this paper, the design problem of multi-UAVs formation control is studied based on the consideration of transmission timedelay. With a Bernoulli random variable describing the randomness of time delays, the Lyapunov-Krasovskii formula is constructed to derive the formation criterion. The multiUAV formation controller is designed on the basis of combining Lyapunov stability theory with stochastic analysis. Finally, a numerical simulation is carried out to verify the effectiveness of the proposed formation controller.

Design of Fuzzy-PID Control Virtual Instrument for 3-RRR Planar Parallel Mechanism

LIU Xia, SHAN Ning, WANG Qing, WU Xinwei

2019, 37(4):  565-572.  doi:10.3969/j.issn.0255-8297.2019.04.013

Asbtract ( 525 )   PDF (5272KB) ( 60 )  

References | Related Articles | Metrics

The application of 3-RRR planar parallel mechanism in engineering field is seriously restricted by the motion precision problem. In order to enhance mechanism motion precision in effect, a fuzzy-PID control system of 3-RRR planar parallel mechanism is established. The experimental control device is set up. And the virtual instrument software of fuzzy-PID real time control system is developed based on LabVIEW. And real time control of mechanism's motion error is studied experimentally. Experimental results show that the motion error of the system can be controlled in real time. It features in arithmetic simplicity and high precision. The developed virtual instrument software not only has a friendly and easy operation interface, but also has high flexibility, scalability, and intelligence.