Table of Content

25 May 2022 Volume 35 Issue 5

  

Improved Image Matching Algorithm Based on LK Optical Flow and Grid Motion Statistics

LIU Qunpo,XI Xiulei,YANG Lingxiao

Electronic Science and Technology. 2022, 35(5):  1-6.  doi:10.16180/j.cnki.issn1007-7820.2022.05.001

Abstract ( 234 )   HTML ( 80 )   PDF (2968KB) ( 133 )  

Figures and Tables | References | Related Articles | Metrics

In order to solve the low accuracy and time-consuming problem of AKAZE algorithm in the image matching of glass-encapsulated electrical connectors, improved image matching algorithm based on LK optical flow and grid motion statistics is proposed in this study. First, the AKAZE algorithm is used to extract feature points, and the M-LDB descriptor is used to describe the features. Then, LK optical flow method is used to calculate the matching area for conditional constraints to obtain matching points, and the FLANN algorithm is adopted for feature matching. Finally, the glass-encapsulated electrical connector image is divided into multiple grids, and the numbers and threshold values of correct matching points of the neighborhood of the feature points that are matched by FLANN are calculated to eliminate the wrong matching points. By using public datasets and glass-encapsulated electrical connector data,the performance of the algorithm is verified and analyzed from both real-time and accuracy aspects. The results show that the improved algorithm has a matching accuracy of more than 93% when dealing with image pairs of glass-encapsulated electrical connectors with blur, brightenss and rotation changes, and the time-consuming is within 0.4 s, which proves the effectiveness of the algorithm.

Research on Supervision Object Detection Based on Improved SSD

HUANG Jing,XIE Xuan

Electronic Science and Technology. 2022, 35(5):  7-13.  doi:10.16180/j.cnki.issn1007-7820.2022.05.002

Abstract ( 159 )   HTML ( 7 )   PDF (2380KB) ( 35 )  

Figures and Tables | References | Related Articles | Metrics

In view of many problems caused by manual acceptance in decoration projects, this study proposes an improved SSD algorithm and applies it to supervision work to replace manual acceptance and promote the realization of intelligent supervision. Because the SSD algorithm has problems such as rechecking the same target and poor detection of small targets, the DPN network is employed to replace the basic feature extraction network VGG16. DPN combines the advantages of Resnet and Densenet, and has better feature extraction capabilities. Feature maps are fused by weighted FPN to highlight the contributions of feature maps of different layers and enrich the semantics of feature maps for prediction. Using depth separable convolution can reduce the amount of model parameters and improve the inference speed of the algorithm. Experimental comparison shows that the average accuracy of the improved model is increased by 3.47%, and the average accuracy of small numbers of detection is increased by up to 15%, which proves that the new model is effective in the task of supervision target detection.

PSO Algorithm Based on Improved Contraction Factor

WANG Pengfei,REN Lijia,GAO Yan

Electronic Science and Technology. 2022, 35(5):  14-19.  doi:10.16180/j.cnki.issn1007-7820.2022.05.003

Abstract ( 474 )   HTML ( 26 )   PDF (657KB) ( 94 )  

Figures and Tables | References | Related Articles | Metrics

The performance of PSO algorithm optimization is affected by the speed update formula. Too fast convergence speed may cause the algorithm to miss the global optimal solution, and the slow convergence speed may cause the algorithm to fall into the local optimal solution. To solve this problem, this study proposes a PSO optimization algorithm based on improved compression factor, namely FPSO. By introducing the compression factor equation, the speed iteration formula is improved, and the influence of the improper setting of the learning factor on the algorithm is reduced. The new adjustment mechanism not only guarantees the convergence performance of the PSO algorithm, but also weakens the influence of the speed boundary on the algorithm. Finally, five classic functions are selected to test the performance of the proposed algorithm. The test results show that compared with the traditional PSO algorithm, the proposed algorithm improves the global convergence ability and shortens the convergence time.

Adaptive Weighted Data Fusion Algorithm Based on Batch Estimation

SHI Zhenhua,ZHANG Na,BAO Xiaoan,SONG Jie

Electronic Science and Technology. 2022, 35(5):  19-25.  doi:10.16180/j.cnki.issn1007-7820.2022.05.004

Abstract ( 418 )   HTML ( 9 )   PDF (1979KB) ( 77 )  

Figures and Tables | References | Related Articles | Metrics

In this study, an adaptive weighted data fusion algorithm based on batch estimation is proposed for multi-sensor data fusion. The algorithm uses time series and spatial sequences to find the variance of the collected data in batches, and uses data consistency detection to eliminate noise, and then obtains the adaptive factors. Subsequently, the adaptive weighting method is used to fuse the data to obtain the predicted value. The simulation experiments with IoT data show that the adaptive weighted multi-sensor data fusion technology of batch estimation can improve the accuracy of sensor measurement and the reliability of the system when processing data, and the adaptive weighted average method based on batch estimation is 10% less than the root mean square error of traditional adaptive method and the accuracy is improved by 2.3%.

Soft-Sensing of Effluent BOD Based on VW-IGRBF Neural Network

ZHAO Doudou,ZHANG Wei

Electronic Science and Technology. 2022, 35(5):  26-32.  doi:10.16180/j.cnki.issn1007-7820.2022.05.005

Abstract ( 153 )   HTML ( 7 )   PDF (1973KB) ( 34 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem that the wastewater treatment process has complex nonlinear characteristics and the effluent BOD is difficult to accurately measure, a soft measurement method based on VW-IGRBF neural network is proposed in this study. The activation function of the neural network is a linear combination of the inverse square root function and the Gaussian function, which makes up for the saturation of a single activation function in certain intervals, and improves the expression and self-adaptability of the hidden activation function. Since the width of the activation function has a greater impact on the generalization performance of the model, a variable width strategy based on kernel density is introduced to improve the generalization ability of the network. In this study, the improved LM algorithm is used to realize the online learning of neural network parameters. Simulation experiments based on actual operating data of the wastewater treatment process show that the proposed VW-IGRBF method has higher prediction accuracy and better adaptive ability for effluent BOD.

Research on Azimuth Error Compensation Based on BP Neural Network at Small-Angle Deviation

DING Huihui,SHAO Tingting,QIAO Xi

Electronic Science and Technology. 2022, 35(5):  33-37.  doi:10.16180/j.cnki.issn1007-7820.2022.05.006

Abstract ( 168 )   HTML ( 8 )   PDF (688KB) ( 38 )  

Figures and Tables | References | Related Articles | Metrics

Deviation angle and azimuth angle are the main measurement parameters in borehole trajectory calculation. However, the measurement error of azimuth angle of the inclinometer with small-angle deviation is larger than that with conventional deviation. In order to improve the accuracy of azimuth measurement of inclinometer under small-angle deviation, the measurement of azimuth angle under 5°~10° is compensated based on BP neural network algorithm. The neural network is established, whose input is two dimensional vector including standard deviation angle and measured azimuth, and output is the expected azimuth. The learning samples are divided into training sets and test sets by random selection, which can make the network have better generalization ability. The simulation results show that the BP neural network error correction model runs stably, with a compensation accuracy of 10^-6, which can increase the measurement accuracy of the low angle of the small angle well deviation from ±5.3° to within ±1.7°.

Fault Diagnosis of Wind Turbine Pitch Actuator Based on Unknown Input Set-Membership Observer

FAN Xiaomin,ZHANG Wei

Electronic Science and Technology. 2022, 35(5):  38-46.  doi:10.16180/j.cnki.issn1007-7820.2022.05.007

Abstract ( 158 )   HTML ( 6 )   PDF (1121KB) ( 26 )  

Figures and Tables | References | Related Articles | Metrics

Wind turbines are generally installed in harsh environments, and their pitch actuators are prone to failure. In view of the pitch actuator failure of a type of wind turbine system with unknown but bounded interference and noise, a collective unknown input observer is designed to detect and separate the pitch actuator failure. The model of wind turbine system is established by using aerodynamic mechanism and modern identification principle and the disturbance is decoupled by optimizing the design of unknown input observer. The interval envelope without considering the fault is obtained by zonotopes, which is used as the upper and lower dynamic threshold of the residual estimation to realize the state estimation. Based on the proposed design, a fault diagnosis strategy using a bank of unknown input observers is presented. The simulation shows that the time and location of the sudden and slow time-varying faults in the 3rd and 5th order linear systems of the pitch actuators are accurately diagnosed by the designed set unknown input observer, which verifies the effectiveness of the proposed fault diagnosis strategy.

Design of DDS Communication Middleware for Resource Limited Devices

DAI Jiangtao,GAO Bo,WAN Jiajun

Electronic Science and Technology. 2022, 35(5):  47-55.  doi:10.16180/j.cnki.issn1007-7820.2022.05.008

Abstract ( 291 )   HTML ( 8 )   PDF (1511KB) ( 50 )  

Figures and Tables | References | Related Articles | Metrics

DDS has good real-time performance, scalability and data processing capabilities, and is suitable for large data volumes and diversified communication requirements in distributed scenarios. In view of the problem that the standard DDS communication middleware is too complex and huge to be deployed on the embedded devices with limited resources, after in-depth analyzing and research on the DDS specification, this study re-designs and optimizes the functions of DDS, and proposes a symmetric publish subscribe mechanism based on Kalman filter model, which reduces the redundant data in communication and realizes a lightweight and portable DDS communication middleware with C language. The communication middleware has been tested and verified on ARM + X86 platform. The experimental results show that the middleware can provide basic DDS services for applications, and improve the response speed of the system and save communication bandwidth resources in the case of occupying a small amount of resources.

Tumor Intelligent Auxiliary Diagnosis Method Based on Machine Learning

CHENG Shunda,CHENG Ying,SUN Shijiang

Electronic Science and Technology. 2022, 35(5):  56-59.  doi:10.16180/j.cnki.issn1007-7820.2022.05.009

Abstract ( 249 )   HTML ( 9 )   PDF (921KB) ( 49 )  

Figures and Tables | References | Related Articles | Metrics

In the field of tumor diagnosis, the artificial intelligence-assisted diagnosis system can accurately distinguish and diagnose tumor attributes and malignant tumor stages, thereby prolonging the survival time of patients. In this study, taking breast tumor as a case, in view of the over-fitting problem caused by the excessive amount of data in the feature extraction process, a supervised learning artificial intelligence-assisted diagnosis model is proposed. When extracting features, hierarchical clustering analysis is introduced to perform effective feature reduction, and the classified feature data is used as the feature input of the artificial neural network model to achieve effective training of the classifier. The experimental results show that compared with other algorithm, the accuracy and AUC value of the proposed algorithm are improved, indicating that the model can not only solve the over-fitting problem caused by the description of massive feature regions, but also enhance the artificial intelligence-assisted diagnosis, thereby completing the mammography target breast tumor high-precision distinction.

Research on Predictive Ammonia Desulfurization Control System

LI Meng,MA Lixin

Electronic Science and Technology. 2022, 35(5):  60-65.  doi:10.16180/j.cnki.issn1007-7820.2022.05.010

Abstract ( 136 )   HTML ( 7 )   PDF (840KB) ( 30 )  

Figures and Tables | References | Related Articles | Metrics

The ammonia desulfurization control system in the coal-fired power generation system has the characteristics of time-varying, large delay and large inertia. Aiming at the typical large delay system, a desulfurization control system based on Smith predictor is proposed to improve the real-time performance of the system by offsetting the pure delay link in the controlled object. The Smith predictor relies on an accurate PID model, and there are still uncertainties in the PID parameter tuning in the system. The SOA algorithm is used to adjust the PID parameters, and the absolute error integral ITAE is used as the fitness function. Through the continuous optimization of the fitness value, the best combination of the proportional integral differential coefficient is found. The simulation results show that compared to the desulfurization system under the control of Smith predictor, the new system has good follow ability and rapid response, the adjustment time is reduced to 9% of the original, and the peak time is reduced to 5% of the original, and has a very small overshoot. The real-time and stability of the system are guaranteed. These results indicate that the algorithm is effective for engineering control systems and has a good prospect of popularization and application.

A High-Precision Synchronization Scheme Design for Short-Term Burst Signals

WAN Jiajun,YUE Chunsheng,SUN Hongsheng,DAI Jiangtao

Electronic Science and Technology. 2022, 35(5):  66-73.  doi:10.16180/j.cnki.issn1007-7820.2022.05.011

Abstract ( 283 )   HTML ( 19 )   PDF (4359KB) ( 93 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems of long acquisition time, high algorithm complexity and high resource overhead in short-term burst communication, a high-precision synchronization scheme is designed. The scheme customizes the frame structure, uses differential matching algorithm of the unique code to achieve signal detection and frame synchronization, and adopts forward estimation algorithm based on pilot sequence to achieve frequency offset estimation. In order to take into account the accuracy of frequency offset estimation and the algorithm complexity, a sliding correlation algorithm is applied for the unique code to achieve initial phase capture, which makes the system compact and more suitable for FPGA hardware implementation. Theoretical analysis and simulation experiments show that the scheme has better receiving performance under the conditions of low signal-to-noise ratio and high dynamic range. The frequency synchronization accuracy can reach one ten thousandth of the symbol rate. Hardware co-simulation and verification are carried out, and the corresponding simulation waveforms are provided in the present study.

Control Strategy of Supercapacitor Based on Train Braking Energy Recovery

ZENG Guang,YANG Jian,SONG Ruigang

Electronic Science and Technology. 2022, 35(5):  74-80.  doi:10.16180/j.cnki.issn1007-7820.2022.05.012

Abstract ( 302 )   HTML ( 8 )   PDF (953KB) ( 51 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem of recovery and utilization of energy feedback from train electric braking, this study analyzes train operation characteristics and focuses on the relationship between train operation process and traction network voltage fluctuations. Combining the dynamic mismatch between the power supplied by the traction substation and the electric power of the train, a vehicle-mounted braking energy recovery and utilization system is adopted, and a super capacitor charging and discharging control strategy based on the dynamic change of the electric power of the train is proposed. The program is mainly composed of super capacitor series-parallel energy storage units and bidirectional DC-DC converters. MATLAB is used for modeling and simulation analysis. The simulation experiment results show that the control strategy can effectively recover and utilize the braking energy of the train while restraining the voltage fluctuation of the traction network, which provides an effective solution for the recovery and utilization of the braking energy of the train.

Parametric Fault Characteristics Analysis of Electrolytic Capacitor in NPC Three Level Inverter

WANG Xin,HUANG Chong,XU Xiang

Electronic Science and Technology. 2022, 35(5):  81-86.  doi:10.16180/j.cnki.issn1007-7820.2022.05.013

Abstract ( 214 )   HTML ( 8 )   PDF (823KB) ( 49 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem that it is difficult to extract the fault features of electrolytic capacitor in NPC three-level inverter, a fault feature extraction method based on variational mode decomposition and modal energy is proposed. By collecting the output current signal of NPC three-level inverter and combining with the reference current signal, the current deviation signal is obtained. According to the characteristics of frequency distribution of current deviation signal, the decomposition scale of VMD is optimized by modal repetition rate. The VMD is employed to decompose the current deviation signal to obtain the limited bandwidth modal component with the center frequency. According to the information entropy of the modal component, the characteristic component which can represent the capacitor fault is determined. Then the modal energy of the characteristic component is calculated, the characteristic vector is constructed, and the characteristic variation rule is sought and classified. The results show that this method can accurately reflect the working state of electrolytic capacitor.