Table of Content

30 September 2023, Volume 41 Issue 5

Previous Issue    Next Issue

Communication Engineering

Optical Fiber Memory Based on Phase Change Material Ge₂Sb₂Te₅

YIN Jiayue, CHENG Siying, LOU Cunkai, YANG Bozhi, ZHANG Yu

2023, 41(5):  727-737.  doi:10.3969/j.issn.0255-8297.2023.05.001

Asbtract ( 144 )   PDF (8717KB) ( 147 )  

References | Related Articles | Metrics

The typical functions of optical fiber are communication and sensing, this paper gives the function of optical fiber storage and designs an all-fiber memory to meet the needs of intelligent development of optical fiber communication systems. In this paper, single-mode fiber (SMF) and multimode fiber (MMF) are used to coaxial soldering, and Ge₂Sb₂Te₅ (GST) material is deposited on the end face of MMF by the magnetron sputtering method, then the end face will emit the Bessel-like beam that can switch the phase state of GST, the length of MMF affects the end face light field, and finally 1.5 mm long MMF is selected to achieve non-volatile memory with arbitrary level access ability, high optical contrast, good stability, and high repeatability. The memory can realize 11 levels of storage randomly and stably, with an optical contrast of 50% and repeated cycles at least 34 times.

Contact Plan Design Based on Data Check Time-Space Graph in Satellite Networks

DAI Cuiqin, HE Liming, LUO Yi

2023, 41(5):  738-752.  doi:10.3969/j.issn.0255-8297.2023.05.002

Asbtract ( 118 )   PDF (808KB) ( 143 )  

References | Related Articles | Metrics

To address network time-space connectivity and traffic congestion issues in low earth orbit satellite networks, a contact plan design scheme based on data check timespace graph (DCTSG-CPD) is proposed via jointly considering the inter-satellite links state, satellite energy efficiency and load state. Firstly, a data check time-space graph (DCTSG) is constructed based on the battery model, and the contact plan design is modeled as a problem of maximizing throughput under constraints of inter-satellite links and satellite energy. Secondly, on the basis of the given DCTSG, the network congestion state is judged through comparing the checkout data and forwarded data. An available contact plan (CP) is generated according to the satellite load and the traffic congestion state of inter satellite link. Finally, the fitness function is designed to evaluate the available CP, and a data check genetic algorithm (DCGA) is developed to update the optimal CP with maximum throughput. Simulation results demonstrate that the proposed DCTSG-CPD scheme improves throughput, and reduces data delivery delay effectively.

Signal and Information Processing

Infrared Dim and Small Target Detection Algorithm Based on Low-Rank and Reweighted Sparse Representation

YANG Yadong, HUANG Shengyi, TAN Yihua

2023, 41(5):  753-765.  doi:10.3969/j.issn.0255-8297.2023.05.003

Asbtract ( 125 )   PDF (2679KB) ( 146 )  

References | Related Articles | Metrics

The detection of infrared dim and small targets is one of the key technologies in the infrared warning system. It remains challenging to accurately, quickly, and robustly detect dim and small targets. This paper proposes an infrared dim and small target detection algorithm based on low-rank and reweighted sparse representation. The algorithm formulates a new optimization equation to more accurately describe the rank of the background matrix and utilizes the structure tensor to extract local prior information. Experimental results show that the proposed algorithm improves the accuracy, speed, and robustness of detecting dim and small infrared targets.

Land Cover Classification of Sentinel-2 Image Based on Multi-feature Convolution Neural Network

HUANG Xianpei, MENG Qingxiang

2023, 41(5):  766-776.  doi:10.3969/j.issn.0255-8297.2023.05.004

Asbtract ( 169 )   PDF (1320KB) ( 185 )  

References | Related Articles | Metrics

The 10 m resolution Sentinel-2 images takes the spectral values of the image as input in the original GoogLeNet without treating the ground objects in the image as a whole. To leverage object-oriented features in Sentinel-2 remote sensing image classification, this paper proposes an Object-oriented GoogLeNet network structure based on multiple features. Object-oriented GoogLeNet incorporates object-oriented spectral and shape features, and fully utilizes the shape features of differences between different ground objects for classification. On the data set of cloudless images in Wuhan and its surrounding areas, the overall accuracy of the classification results of Object-oriented GoogLeNet model has increased by 1.773% compared to GoogLeNet. The results show that the model with object-oriented features enhances the classification performance of Sentinel-2 remote sensing images.

Sparse Set Object Detection Combined with Transformer Multi-scale Instance Interaction

KAN Yaya, ZHANG Sunjie, XIONG Juan, ZU Yi

2023, 41(5):  777-788.  doi:10.3969/j.issn.0255-8297.2023.05.005

Asbtract ( 120 )   PDF (9049KB) ( 172 )  

References | Related Articles | Metrics

In order to improve the problem of lack of spatial detail information in feature maps, failure of target features to interact with global context instance, and insufficient learning of global semantic information, a sparse set object detection algorithm combining adaptive feature augmentation and instance feature interaction is designed. In the process of feature extraction, the adaptive feature augmentation module uses pooling and convolution at different scales to enrich high-level semantic information, and reduces noise interference such as the low-level semantic information background. Meanwhile, it decreases the rate of false detection and missed detection. In design of bounding box regression, the instance feature interaction module combines multi-layer attention of transformer which enhances the channel information of the proposal box. Channel attention and dynamic convolution network are also employed to improve the edge information of the object and increase the interaction efficiency of the network instance feature. Finally, experiment results show that the average accuracy of COCO2017 dataset is improved by 4.2%, 4.6% on the large target, and 2.7% on PASCAL VOC dataset, respectively.

Void Filling of DEM in a Generative Adversarial Network Fused with Self-Attention Mechanism

ZHANG Chunsen, ZHU Jiangle, ZHANG Xuefen, LIU Xudong, SHI Shu

2023, 41(5):  789-800.  doi:10.3969/j.issn.0255-8297.2023.05.006

Asbtract ( 123 )   PDF (10129KB) ( 163 )  

References | Related Articles | Metrics

Aiming at the problems of existing DEM data void filling algorithms, such as discontinuous repair effect, narrow applicable null value range and loss of detail reconstruction, this paper proposes a DEM void filling method integrating self-attention mechanism with generative adjunctive network. Firstly, a self-attention mechanism is constructed to extract DEM data feature information to improve the elevation discontinuity and texture detail loss of DEM cavity filling results. Secondly, symmetric convolutional and deconvolution network structures are used in the generator to ensure the generation of high reliability data to realize the filling of the void region, and the discriminator is used to realize the pre-classification of the filling results. Finally, combined with the reconstruction of loss function and the generation of adversarial loss function, the network training was carried out to improve the robustness of DEM cavity filling results to outliers and enhance the regression ability of the model. The experimental results show that compared with the filling results of spatial interpolation and deep learning, the proposed method can greatly improve the filling accuracy and effectively solve the problems of holes in the original data.

High-Quality Urban Photos Collection Method Based on Human-Machine Cooperation

CHEN Huihui, ZHONG Weizhao

2023, 41(5):  801-814.  doi:10.3969/j.issn.0255-8297.2023.05.007

Asbtract ( 100 )   PDF (1780KB) ( 93 )  

References | Related Articles | Metrics

Many smart city applications require various photos collected in urbans. In this paper, contexts of sensing-task photos collected by people are converted into task instruction to drive the robot to collect more urban photos. The influence of parameters such as distance between people and objects, person’s photographing posture, and land slope on sensing effectiveness is evaluated to select appropriate task parameter values. To address deviation of task command conversion during human-machine cooperation, we present approximate nearest neighbor with dynamic threshold (DTANN) algorithm and propose two methods for optimizing task command parameters including camera posture and robot positions. The overlap of photos collected by humans and machines is used as the quality indicator to evaluate the tasks completion of the robot. Experimental results show that with the aid of DTANN method and task instruction optimization, the precision rate and recall rate both have been improved, and the F1-measure value is increased by 8% to 20%.

Singing Voice Separation Method of Unet Based on Squeeze-and-Excitation Residual Group Dilated Convolution and Dense Linear Gate

ZHANG Tianqi, XIONG Tian, WU Chao, WEN Bin

2023, 41(5):  815-830.  doi:10.3969/j.issn.0255-8297.2023.05.008

Asbtract ( 147 )   PDF (6107KB) ( 175 )  

References | Related Articles | Metrics

To improve speech timing information capture and utilize underlying features in Unet frequency domain singing voice separation network model, a convolutional neural network with smaller parameters and better song separation effect is proposed in this paper. Firstly, a residual group dilated convolution combined with squeeze-and-excitation module is incorporated into the encoding and decoding stage. While reducing the number of parameters and increasing the receptive field of the network, it can adaptively learn the importance of different channel features, so as to enhance the useful features and suppress the irrelevant ones. Secondly, in the transmission layer, the gating linear units are connected by dense addition to enhance the acquisition of temporal features in the process of feature transmission, and the dilated convolution is used to replace the ordinary convolution to expand the receptive field of the network. Finally, the attention gating mechanism is used to replace the jump connection in the baseline Unet to enhance the utilization of the underlying features. Experiments were conducted on the Ccmixter and MUSDB18 datasets, compared with the baseline network, the proposed approach achieves improvement in voice separation performance with only about one-fifth of the parameters.

Implementation and Acceleration of Linear KNN Algorithm for Laser Point Cloud Based on FPGA

CHEN Xiaoyu, YANG Mengxue, LI Changdui, ZHAO Pengcheng

2023, 41(5):  831-839.  doi:10.3969/j.issn.0255-8297.2023.05.009

Asbtract ( 165 )   PDF (1729KB) ( 195 )  

References | Related Articles | Metrics

To address the time-consuming problem of 3D laser point cloud for linear K-nearest neighbor (KNN) search, a fast KNN search method based on multi-processor system on chip (MPSoC) field-programmable gate array (FPGA) is proposed. Firstly, the implementation framework of 3D laser point cloud KNN algorithm based on MPSoC FPGA is given. Then, the design ideas and implementation process of each module are elaborated. Finally, the proposed method is validated through tests and verification on platform built based on MZU15A development board and TM-LIDAR-16. Results demonstrate that the 3D laser point cloud KNN algorithm based on MPSoC FPGA can reduce time consumption while ensuring the accuracy of neighboring point search.

Computer Science and Applications

Research on Pipe Sorting Based on Improved Mask R-CNN Using Feature Fusion and Region Generation Network

HAN Huiyan, WU Weizhou, WANG Wenjun, HAN Xie

2023, 41(5):  840-854.  doi:10.3969/j.issn.0255-8297.2023.05.010

Asbtract ( 127 )   PDF (13576KB) ( 53 )  

References | Related Articles | Metrics

In order to solve the problems of difficulty in distinguishing various pipe fittings and interference of light and shadow on segmentation accuracy, a pipe fittings sorting algorithm based on improved Mask R-CNN is proposed. The feature fusion network is improved by incorporating low-level feature map, improving the recognition rate of small pipe fittings. The generation box of area growth network is modified according to the size ratio of pipe fitting, so to accelerate the convergence rate of the model. Introduction of the channel and space attention module enhances the identification accuracy of pipe fitting and mask effect. The improved Mask R-CNN is applied to the sorting task of four types of pipe fittings, demonstrating increased mAP and mRecall values for mask detection (1.5% and 1.7% improvement, respectively). The robot exhibits enhanced capabilities in discriminating the location, type and size of pipe fittings, thereby meeting the accuracy requirements of sorting pipe fittings in actual production.

Attack Modeling Combined with Industrial Control Operati

ZHANG Yaofang, ZHANG Zheyu, LI Tongtong, SUN Jun, WANG Zibo, WANG Bailing

2023, 41(5):  855-869.  doi:10.3969/j.issn.0255-8297.2023.05.011

Asbtract ( 103 )   PDF (1860KB) ( 73 )  

References | Related Articles | Metrics

Existing industrial environment attack graph lacks the association display of protocol vulnerabilities, attack semantics and control operations, resulting in missing representations of the intelligent manufacturing system attack graph. The impact of system process operations on the underlying equipment cannot be described in the graph. Therefore, this paper proposes the attack graph association rules for proprietary protocols and specific attacks in industrial environments. Attack graphs for intelligent manufacturing systems with extended semantics are generated based on search algorithms. Furthermore, a three-layer attack graph model incorporating vulnerability, host, and operation layers is designed to integrate business operations into the attack graph for correlated display. Experimental results show that the extended rules and model can effectively enrich and describe the multi-step attack process of the intelligent manufacturing system.

Constructing Sentiment Lexicon in the Education Field by Integrating Skip-Gram and R-SOPMI

CHEN Jun, XI Ningli, LI Jiamin, WAN Xiaorong

2023, 41(5):  870-880.  doi:10.3969/j.issn.0255-8297.2023.05.012

Asbtract ( 126 )   PDF (710KB) ( 163 )  

References | Related Articles | Metrics

This paper presents a method for constructing a fine-grained Sentiment Lexicon in Education to address specific emotional issues in sentiment analysis of educational feedback texts. First, we construct an educational domain corpus, which contains emotional features in both formal and informal domains. Second, a fusion-based method is proposed to construct a domain Sentiment Lexicon by identifying linguistic probability features and statistical probability features of words through sentiment classification. The proposed repetitive semantic orientation pointwise mutual information (R-SOPMI) algorithm enhances SO-PMI for sentiment classification, enabling co-occurrence multi-category sentiment classification. Finally, a fine-grained Sentiment Lexicon in the field of education is obtained, and the dictionary expands to 39 138 emotional words. Experiment results show that except for “anger”, the F1 of the emotion category of the constructed educational field emotion dictionary is all higher than 78.09%. Compared with a general dictionary, the Macro_Precision, Macro_Recall and Macro_F1 increased by 21.95%, 2.50% and 13.01%, respectively. The fusion feature method effectively extracts domain features, facilitating the construction of a comprehensive fine-grained domain dictionary.

Dual Authorization Sharing Scheme of Searchable Electronic Medical Data Based on Consortium Blockchain

MA Xue, PAN Heng, YAO Zhongyuan, SI Xueming

2023, 41(5):  881-895.  doi:10.3969/j.issn.0255-8297.2023.05.013

Asbtract ( 125 )   PDF (816KB) ( 130 )  

References | Related Articles | Metrics

Retrieval of electronic medical record (EMR) in cloud environments induces security problems and patient privacy data leakage problems. To this end, a dual-authorization sharing scheme for EMR that supports on-chain keyword ciphertext retrieval is proposed. In the scheme, original medical data ciphertexts are stored in a cloud, and the information of medical data keyword index is constructed with searchable encryption technology and stored on the blockchain. On the premise of obtaining the hospital retrieval authority, a keyword retrieval algorithm under distributed conditions is used to realize the secure re trieval of the medical data ciphertexts. Based on searchable proxy re-encryption algorithm, an authorization on-chain method for the electronic medical data is proposed, which en sures the access control of patients’ medical data and realizes a double authorization of the shared medical data by the hospital and patients. Finally, random oracle model is used to verify the semantic security of the scheme under the assumption of n-QBDH, and the superiority of the scheme in terms of computational cost is proved by experiments.

Charge-Discharge Optimization for Shared Electric Vehicles Under Carbon Trading Regulation

LI Junxiang, HE Wenting, WANG Jinling

2023, 41(5):  896-910.  doi:10.3969/j.issn.0255-8297.2023.05.014

Asbtract ( 82 )   PDF (1701KB) ( 180 )  

References | Related Articles | Metrics

In this paper, a charge-discharge optimization model of electric vehicles participating in carbon trading market and peak-shaving auxiliary service market is proposed for the study of shared electric vehicles under the switching mode. In the upper model, EV charge and discharge scheduling is carried out with the goal of minimizing the daily operating cost of EV operators, while the lower model continues to optimize the scheduling results of the upper layer with the goal of minimizing the fluctuation of power grid load. Then, this paper compared and analyzed the charge-discharge schemes from the perspectives of EV operators and the power grid. The carbon emissions produced by electric vehicle operators are compared with that by gas-powered vehicle operators over one operating cycle. Finally, simulation results show that the proposed model can meet users’ travel needs while reducing the cost of operators and the load fluctuation of power grid.