Computer and Modernization

Skinned Mesh Generation Algorithm with Boundary Adaptive Ability

SHU Pei-tong, ZHAO Jia-bao

2023, 0(02): 1-5.

Asbtract ( 194 )

PDF (2912KB) ( 51 )

References | Related Articles | Metrics

Mesh generating from the original image plays a key role in skinned mesh animation. The quality of the mesh directly affects the animation quality. However， mesh optimization is time-consuming and greatly dependent on the producer’s skills. This paper presents a skin mesh generation algorithm that can adapt to the image directly. Firstly， the uniform approximate polygon surrounding the image is calculated using the image contour’s normal vector transformation. Then， by using the information entropy as the weight to measure the internal flexibility of the image and combining an improved centroid Voronoi algorithm， the internal sampled points are obtained. Finally， the Delaunay triangulation with boundary constraints is performed by combining the vertices of the boundary polygon with the internal points to obtain a triangular mesh for skinned mesh animation. The experimental results show that the proposed algorithm can generate high-quality meshes， which perfectly match the original image boundary and can be used for skinned mesh animation.

Fast Path Planning Algorithm in 3D Space for UAV

SHEN Hao-yang, CHEN Xiao-lei, YUAN Jun-ling, HAN Lu

2023, 0(02): 6-11.

Asbtract ( 188 )

PDF (1872KB) ( 33 )

References | Related Articles | Metrics

Most of the path planning algorithms used in the path planning of unmanned aerial vehicles such as UAVs and unmanned underwater robots working in three-dimensional space are RRT* algorithms. However， the RRT* algorithm has problems such as slow convergence speed， large number of iterations， and frequent collision detection. As a result， the path planning ability of UAVs is poor and cannot meet the application needs of many scenarios. Aiming at the insufficiency of existing algorithm sampling. Bidirectional-self-optimizing growing tree（B-SOGT*） was proposed， which used bidirectional search， double gravitational field， and tentative elastic expansion to realize path planning in three-dimensional space. The two-way search takes the starting point and the target point as the root nodes respectively， and constructs 2 random trees to perform spatial search at the same time， which improves the search efficiency； The double gravitational field is a gravitational field generated by the root nodes of 2 random trees respectively， and the sampling efficiency is improved under the action of the gravitational field; The tentative elastic expansion method introduces the parent node re-selection mechanism when generating the path， which does not require collision detection and improves the calculation speed of the algorithm. Simulation results show that the B-SOGT* algorithm has the advantages of faster convergence， better path quality and fewer iterations.

Outlier Elimination of Telemetry Data Based on 53H Method

SUN Hao, LIU Lei, SUN Jian-wei, DUAN Si-jia

2023, 0(02): 12-16.

Asbtract ( 185 )

PDF (965KB) ( 33 )

References | Related Articles | Metrics

During the distribution of telemetry data from the satellite to telemetry stations，due to the influence of weather，magnetic field interference and other accidental factors，the observed data will seriously deviate from the target true value，which we call this part of data as outliers. The accuracy of telemetry data is very important. If there are outliers in telemetry data，it will directly affect the subsequent fault detection and prediction，resulting in abnormal misjudgment，and even false alarm，which will cause trouble for subsequent staff to confirm satellite status and deal with satellite abnormalities. Based on the traditional 53h method，this paper adds a three-point equal weight head and tail smoothing algorithm to avoid the loss of four data points caused by the traditional 53h method; The judgment method of speckle outliers is added，and the disadvantage of weak ability of 53h method to eliminate speckle outliers is improved; The secondary judgment of adjacent normal data of outlier data is added to avoid the wrong elimination of normal data due to the influence of outliers，so as to ensure the integrity of data. In this paper，the telemetry data of a navigation satellite is used for experiments. Under the same experimental conditions，the same simulation data are used to carry out comparative experiments with the Nair criterion and the least square method，and the characteristics of this algorithm and other outlier elimination algorithms are summarized. The experimental results show that the algorithm can eliminate the outlier data accurately and retain the abnormal data due to satellite faults and anomalies. The algorithm proposed in this paper is simple to implement and fast to eliminate. It can be well applied to the data preprocessing of satellite fault detection based on telemetry data.

Laser SLAM Mapping Method Based on Visual Information

WU Song-lin, ZHANG Guo-wei, LU Qiu-hong, SHI Jian-zhuang, HUANG Wei

2023, 0(02): 17-23.

Asbtract ( 125 )

PDF (1752KB) ( 30 )

References | Related Articles | Metrics

In view of the fact that the Gmapping SLAM algorithm has high requirements on the accuracy of odometry positioning information in the process of map construction， and there are problems such as particle dissipation and degradation， Firstly， a parallel visual recognition and localization network is designed to strengthen the localization ability and improve the semantic information and composition accuracy； Secondly， the optimization proposal distribution is improved， we use the laser observation model to replace the odometrg motion model and perform Gaussian sampling to cover the probability distribution of the robot with fewer particles； Finally， the visual information and laser information are fused by Bayesian rule， and the original algorithm is compared. The results show that the algorithm improves the accuracy and robustness of map construction and enriches the semantic information.

Latency Optimization of Distributed Key-value Store Based on RDMA

WANG Zhe, WANG Yu-mei, WU Ya-fei, ZANG Yi-hua

2023, 0(02): 24-27.

Asbtract ( 222 )

PDF (1009KB) ( 47 )

References | Related Articles | Metrics

Distributed key-value store is widely used in industry， and methods to improve the performance of data transfer in distributed key-value store has drawn great interests in the study of distributed systems and parallel computing. RDMA （Remote Memory Direct Access） is a popular network technology in high-performance computing. Making use of its low latency and high bandwidth features can lead to enormous improvement of the performance of distributed key-value store. MPI offers routines to perform remote memory access （RMA）， also known as one-sided communication， which can benefit from RDMA technology. In this paper， we propose leverageing MPI one-sided communication refactors key-value operations and redesigns the communication model according to the features of one-sided communication. To solve the data inconsistency problem brought by RMA， we modify the storage structure of key-value pair and schedule the communication strategy. We evaluate the improvement by comparing the latency of RMA based key-value operations and TCP/IP based key-value operations when transferring key-value pairs of different lengths and get positive result.

Communication Fault Diagnosis Knowledge Base Construction and Iteration Based on Semi-supervised Clustering

HONG Tao, ZHU Peng-yu, GUO Bo, WANG Jing-Yu

2023, 0(02): 28-33.

Asbtract ( 104 )

PDF (1076KB) ( 29 )

References | Related Articles | Metrics

Power communication network is an indispensable and important part of power system， it’s the basis of power grid dispatching automation and production management modernization， and an important technical means to ensure the safe， economic and stable operation of power grid. The diagnosis of communication faults still depends on manual experience， which is difficult to meet the safety production needs of increasingly large and complex communication network. Methods based on rule engine or neural network gradually encounter bottlenecks in the application of production environment. It is difficult to train due to less samples， or work independently in production environment as a black box. To solve the above problems， this paper proposes an alarm clustering algorithm based on improved Markov-clustering and a fault diagnosis algorithm based on sequence similarity and OPTICS clustering， adapt to the current small sample scenario of fault data. On the basis of the above algorithm results， the fault diagnosis knowledge base and its iterative learning mechanism are constructed by using a small number of labels. It is verified by the data accumulated in actual production. The results show that the relevant algorithms and knowledge base have a good effect in dealing with actual production problems.

Monocular Depth and Pose Estimation Based on Conditionally Convolution and Polarized Self-attention

QIAO Shan-bao, GAO Yong-bin, HUANG Bo, YU Wen-jun

2023, 0(02): 34-39.

Asbtract ( 121 )

PDF (2458KB) ( 26 )

References | Related Articles | Metrics

This paper proposed a novel monocular depth and pose estimation framework based on view synthesis and the self-supervised structure from motion paradigm by introducing conditionally convolution and polarized self-attention. Conditional convolution assigns multiple groups of dynamic weights to different input data， and all weights share one convolution operation after integration， which improves the model capacity without significantly increasing the computational cost. The image information integrity has significant impacts on the performance of depth estimation tasks. Polarized self-attention keeps the high resolution of data in channel or spatial dimensions through polarization filtering， which could prevent the loss of fine-grained and structural information. The dimension orthogonal to the channel or space is compressed to reduce the computation， and the feature intensity range lost in the compression process is enhanced and dynamically mapped through nonlinear functions. The self-attention mechanism can realize long-distance modeling of data in various dimensions. Experiments on the KITTI dataset demonstrate that the proposed model has excellent performance in self-supervised monocular depth and pose estimation tasks.

A Review of Deep Neural Networks Combined with Attention Mechanism

HUANGFU Xiao-ying, QIAN Hui-min, HUANG Min

2023, 0(02): 40-49.

Asbtract ( 611 )

PDF (2408KB) ( 88 )

References | Related Articles | Metrics

Attention mechanism has become one of the research hotspots in improving the learning ability of deep neural network. In view of the wide attention paid to the attention mechanism， this paper aims to give a comprehensive analysis and elaboration of attention mechanism in deep neural network from three aspects: the classification of attention mechanism， the way of combining with deep neural network， and the specific applications in natural language processing and computer vision. Specifically， attention mechanism has been divided into soft attention mechanism， hard attention mechanism and self-attention mechanism， and their advantages and disadvantages are compared. Then， the common ways of combining attention mechanism in recursive neural network and convolutional neural network are discussed respectively， and the representative model structures of each way are given. After that， the applications of attention mechanism in natural language processing and computer vision are illustrated. Finally， several future developments of attention mechanism are illustrated expecting to provide clues and directions for subsequent researches.

Multi-weather Vehicle Detection Algorithm Based on Modified Knowledge Distillation

CHEN Zhuo, QIAO Gui-fang, CHAI Xin-bo, DU Yi-jun, SHEN Chong-lin, WANG Yuan-hao

2023, 0(02): 50-57.

Asbtract ( 138 )

PDF (8455KB) ( 28 )

References | Related Articles | Metrics

In order to improve the vehicle detection result under multi-weather conditions， a convolutional network based on modified knowledge distillation method was proposed. The network uses cumbersome CNN（Convolutional neural network） as teacher network and lightweight CNN as student network. Without adding new training dataset and slightly increasing the number of light network parameters， the performance of the lightweight CNN under multi-weather vehicle detecting conditions can be improved. The network utilizes a specialized data enhancing method to generate a multi-weather feature dataset. The teacher network is trained on the data without weather feature， and the student network data is trained simultaneously on the data with weather features. Considering that images without weather features can provide more information relatively， through this training method， the student network can better learn the information generated by the teacher network. Finally， through the multi-weather vehicle detecting performance of the network trained and tested on BDD100k dataset with enhanced weather dataset， the detectability and stability of the student model in multi-weather environment boosts. The comparison test of the generalization ability of multiple networks is carried out on DAWN multi-weather dataset， and the modified distillation convolutional network achieves certain advantages in average precision and detection speed.

A Text Entity Linking Method Based on BERT

XIE Shi-chao, HUANG Wei, REN Xiang-hui

2023, 0(02): 58-61.

Asbtract ( 208 )

PDF (1033KB) ( 49 )

References | Related Articles | Metrics

Entity linking is not only an important means to clarify the entity reference in the text， but also the key technology to construct the knowledge map. It plays an important role in the fields of intelligent question answering and information retrieval. However， due to the problems of polysemy or polysemy in Chinese Texts， the accuracy of the existing text entity linking methods is low. To solve these problems， this paper proposes a text entity linking method based on BERT （Bidirectional Encoder Representations from Transformers）， named STELM model. By inputting each pair of reference and candidate entities into the BERT model， the output results are spliced together and the candidate entity with the highest score is taken as the final result through a full connection layer. The experimental results on CCKS2020（2020 China Conference on Knowledge Graph and Semantic Computing） dataset show that the accuracy of the model proposed in this paper has a certain improvement compared with other models and the accuracy has reached 0.9175.

Apples Recognition in Natural Environment Based on Faster-RCNN

SHI Zhan-kun, YANG Feng, HAN Jian-ning, GUO Xin, CAO Shang-bin

2023, 0(02): 62-65.

Asbtract ( 163 )

PDF (1291KB) ( 34 )

References | Related Articles | Metrics

Aiming at the problems of overlapping fruits， interference of branches and leaves， and complex backgrounds in apple orchards， the Faster-RCNN algorithm was proposed. By adding Mosaic data enhancement at the input end， the amount of data is increased and the ability to recognize small objects is enhanced. At the same time， the anchor frame in the Faster-RCNN algorithm is optimized， and the optimized anchor frame can better detect the distance. The target fruit far from the camera and the Soft NMS algorithm are used to further improve the recognition effect of dense areas. The verification results show that the recall rate is 91.44%， the accuracy rate is 93.35%， the F1 value is 92.38%， and the average detection speed per image can reach 0.2 fps. The robustness of the improved algorithm is enhanced， which can meet the recognition of apple fruits in natural environment.

Text Summarization Generation Model Based on PGN-CL

LIU Ya-qing, ZHANG Hai-jun, LIANG Ke-jin, ZHANG Yu, WANG Yue-yang

2023, 0(02): 66-71.

Asbtract ( 151 )

PDF (1952KB) ( 22 )

References | Related Articles | Metrics

The model of abstractive text summarization based on the Seq2Seq framework has made great achievements. However, most of these models suffer from out-of-vocabulary， generated text repetition， and exposure bias. To tackle this problem, we propose a pointer generator network based on adversarial perturbation contrastive learning （PGN-CL） to model the text summarization generation process. As the basic structure， PGN is used for solving the problems of out-of-vocabulary and generated text repetition in this model as well as introducing Adversarial Perturbation Contrastive Learning as a new model training method to address exposure bias. In the model training process， we add perturbations to the target sequence and build a contrastive loss function to generate adversarial positive and negative samples. By this way， negative samples are similar to the target sequence in the embedding space but have large differences in semantic space， while the positive samples are similar to the target sequence in semantic space but have large differences in embedding space. These indistinguishable positive and negative samples can guide the model to learn the distinguishing features of these samples better in the feature space and obtain more accurate summary representation. The experiment result on the LCSTS dataset shows that the proposed model outperforms the comparative baselines on the ROUGE evaluation metric, demonstrating the effectiveness of the proposed model for summary quality improvement.

Mask-wearing Face Recognition Method Fused with Dual Attention Mechanism

SHENG Jiang-an, CHEN Shu-rong

2023, 0(02): 72-77.

Asbtract ( 101 )

PDF (2261KB) ( 36 )

References | Related Articles | Metrics

To address the problem that existing face recognition models cannot effectively extract regional features from faces wearing masks， a face recognition model incorporating a dual attention mechanism is proposed for faces wearing masks. Firstly， a self-constructed face image wearing a mask is used as input data， and ResNet50 is used as the benchmark network to introduce coordinate attention and split attention mechanisms into the residual blocks， where coordinate attention is used to reduce feature extraction in the mask region and reduce feature interference in the mask region； Split attention is used to extract non-mask region features at a fine granularity and extract more features from key areas. The ArcFace classification function is then used to optimize the classification boundary， combined with a cross-entropy loss function as a constraint， to achieve fine-grained recognition of faces wearing masks. The experimental results show that the model in this paper achieves 95.2% recognition accuracy in the test set， which is 1 percent point and 1.5 percent point higher than that of ResNet50 and AttentionNet models respectively.

Multi-target Detection of Transmission Lines Based on Improved YOLOv5

TANG Hao-wei, YAO Jun-cai, YAO Cong-ying, SUN Ying, PEI Xing-yi, SONG Chun-xiao

2023, 0(02): 78-82.

Asbtract ( 227 )

PDF (2417KB) ( 47 )

References | Related Articles | Metrics

In order to realize the identification and detection of transmission line components， a multi-target detection algorithm for transmission lines based on improved YOLOv5 is proposed for the current problems of deepening the number of target detection network layers， increasing the number of parameters and computation， resulting in poor real-time performance. Firstly， the number of parameters in the network was reduced by using the shuffleNetv2 structure as the backbone structure for network feature extraction. Secondly， the BottleneckCSP in the PANet network is changed to a Light_CSP module to speed up feature fusion. Thirdly， the CIoU loss， DIoU-NMS method is used to reduce the loss of position of the prediction frame and the problem of missed detection. Finally， in order to verify the effectiveness of the proposed algorithm， a transmission line image dataset was used for training and testing The results show that the improved YOLOv5 has a parametric count of 7.5×106， a floating point computation of 10.9， an average accuracy of 87.5% and an FPS of 69.2， which meets the requirements for accuracy， lightness and real-time inspection of transmission line components.

A Semi-supervised Model with Non-negative Matrix Factorization for Multiplex Network Clustering

LIU Xing-jian, YANG Xiao-fu, HU Lei

2023, 0(02): 83-88.

Asbtract ( 112 )

PDF (2089KB) ( 27 )

References | Related Articles | Metrics

Real-world multiplex networks often have the characteristics of multi-dimensional and high complexity. The clustering accuracy of existing approaches that only use network topology information for clustering often cannot be guaranteed. To address the problem， the paper proposes a semi-supervised model with non-negative matrix factorization （SeNMF）. Firstly， the model designs a greedy search method based on the PageRank algorithm to obtain the consensus prior information of network. The prior information is used to enhance the topology of each network layer to reduce network noise. Then， the overall non-negative matrix factorization is used to obtain a better common low-dimensional representation matrix by fusing the low-dimensional representations of all network layers on the Grassmannian manifold. Finally， K-means is used to obtain the public community structure of the network. Extensive experiments show that SeNMF achieves the outstanding performance over the state-of-the-art approaches， whether it is the increase of network layers or the enhancement of inter-layer noise.

Research on Network Asset Detection Technology of Industrial Control System

JIANG Xing-yu, XU Rui, ZHANG Ruo-yu, ZHANG Zhi-yong,

2023, 0(02): 89-95.

Asbtract ( 268 )

PDF (1088KB) ( 60 )

References | Related Articles | Metrics

The security of the industrial control system is related to the national economy and people’s livelihood， and is an important part of national security. With the continuous development of the Internet of Things technology， the industrial control system network has penetrated into various industries. However， due to design defects or lack of security means， the relevant assets of the industrial control system are extremely vulnerable to hackers and exploits. Detecting and knowing the industrial control assets exposed to the Internet environment is an important step to realize the information monitoring of the industrial control system， find security loopholes and grasp the security situation of cyberspace. This paper introduces the commonly used detection methods for industrial control system network asset detection. The port detection technology is used to scan the ports on the target host， and then the industrial control protocol and general protocol network asset detection technology is used to discover industrial control equipment and collect asset information according to the port opening. Through the Internet experiment， the data of the detection results are comprehensively analyzed， the characteristics of the network asset detection technology of the industrial control system are summarized， the problems existing in the current technology are pointed out， and the future development is prospected.

A Trusted Computing Based Secure Scheme in Vehicular Named Data Networking

FAN Na, ZOU Xiao-min, LI Si-rui, YANG Xiao-duo

2023, 0(02): 96-103.

Asbtract ( 105 )

PDF (2509KB) ( 36 )

References | Related Articles | Metrics

Named data networking is a content-centered network. Applying named data networking to the Internet of vehicles to form a vehicular named data networking can effectively solve the limitations brought by traditional TCP/IP communication in the Internet of vehicles. However， vehicular named data network also faces security risks， especially interest flooding attack and on-off attack， which seriously affect the information sharing and secure communication of vehicular named data networking. Aiming at the above attacks， this paper proposed a trusted computing based secure scheme in vehicular named data networking called TSSRA. Firstly， by analyzing the impact of attack behavior on the network， the characteristic values identifying malicious behavior are extracted， and then a trusted computing method based on characteristic values is designed to separate malicious behavior from legitimate behavior by evaluating the trust value of nodes. Simulation results show that the secure mechanism proposed in this paper can effectively suppress the malicious behavior of malicious nodes， enhance the security of the network and ensure the safe and efficient operation of the network.

Security Authentication Selection Mechanism for Resource-constrained NB-IoT Nodes

LI Wei-qun, CHANG Chao-wen, LI Peng-jing

2023, 0(02): 104-109.

Asbtract ( 135 )

PDF (1273KB) ( 30 )

References | Related Articles | Metrics

Some narrowband IoT devices cannot use group authentication to access the core network due to resource constraints. Under the framework of 5G network access authentication， a large number of devices accessing the core network at the same time will cause the devices to wait in a queue and cause network congestion. Based on the RFC7228 standard on resource-constrained devices issued by the IETF working group， this paper proposes a group authentication scheme for 3 device types. Firstly， the security and congestion problems of 5G access authentication for narrowband IoT devices are pointed out; then， a solution group authentication scheme is proposed; finally， the access authentication methods are determined for three resource-constrained devices. Experimental results and performance and security analysis show that the proposed scheme meets the resource capabilities required for device group authentication. Compared with the 5G access authentication scheme， the proposed scheme can reduce the number of network signaling by more than 60%， and it decreases with the increase of devices in the group. In terms of security， it has the ability to resist replay attacks， man-in-the-middle attacks and Dos attacks.

Ethereum-based Medical Healthy Data Sharing System

JIANG Yu-hang, ZHANG Xin-you

2023, 0(02): 110-115.

Asbtract ( 137 )

PDF (1755KB) ( 30 )

References | Related Articles | Metrics

The sharing of medical healthy data is of great significance to the supervision of national medical insurance funds， the settlement of medical disputes and medical research. The medical healthy data sharing platform should ensure the authenticity and privacy of relevant data to avoid data leakage and tampering. At the same time， patients’ ownership of their own medical healthy data should be paid enough attention. This paper designs a medical healthy data sharing system based on Ethereum blockchain， which uses blockchain technology to ensure the transparency and immutability of relevant information. At the same time， IPFS is used to cover the shortage of high storage pressure of blockchain to improve the system performance. For another， CP-ABE encryption and ring signature are used to ensure the security of data and the privacy of user identity information. The test shows that the system provides perfect related functions for different user roles and patients have the control of their own data. In the end， it can provide better data security and user identity information privacy.

Portable Dual-channel Digital Oscilloscope in IOT

CHENG Yong-jian, DENG Ling-qi, TAN Jia-jun, YANG Fan, CHEN Liang-liang

2023, 0(02): 116-120.

Asbtract ( 156 )

PDF (1690KB) ( 25 )

References | Related Articles | Metrics

As the “eyes of electronic engineers”， oscilloscope is one of the most commonly used measuring instruments in engineering field. Portable oscilloscope overcomes the disadvantages of large volume and high power consumption of traditional digital oscilloscope. Under the condition of retaining the basic functions of traditional digital oscilloscope， it makes the oscilloscope small and portable， especially suitable for outdoor and the environment with strict requirements on the volume of measuring instruments.This paper designs a portable dual channel digital oscilloscope based on the basic principle of oscilloscope， with FPGA as the core control and combined with the related technology of IOT. The whole design is divided into four modules: MCU main control module， signal sampling module， display module and remote control module. The oscilloscope designed in this paper can accurately measure the signals with frequency range of 0.1 Hz~2 MHz and amplitude range of -5 V~5 V， and can send the data back to the mobile terminal in real time. The test results show that all indicators of the oscilloscope meet the requirements and achieve the expected goal.

A Relay-based Consortium Chain Cross-chain Communication System

FENG Yun-xia, CHEN Hong-da, NIU Yun-he

2023, 0(02): 121-126.

Asbtract ( 147 )

PDF (982KB) ( 33 )

References | Related Articles | Metrics

By studying the difficulty of cross-chain communication between blockchains， this paper develops and improves functions based on the relay-based cross-chain communication mechanism， and designs a relay-based alliance chain cross-chain communication system. Through the cross-chain system， cross-chain requirements such as interconnection of heterogeneous chains， addressing and forwarding of cross-chain requirements， cross-chain resource invocation， and cross-chain contract initiation are realized. A two-phase execution protocol for cross-chain communication is proposed， which fully guarantees the atomicity of cross-chain interaction. The block header synchronization contract is designed for cross-chain transaction verification， which simplifies the cross-chain transaction verification process and reduces the complexity of cross-chain communication. Through the functional test of the experimental simulation system， the cross-chain system designed in this paper can safely and efficiently realize the cross-chain communication operation between the blockchains.

Table of Content