Computer and Modernization

Optimization Method of Hadoop File Archiving Based on LZO

ZHANG Jun, SU Wen-hao

2023, 0(06): 1-6. doi:10.3969/j.issn.1006-2475.2023.06.001

Asbtract ( 4 )

PDF (2368KB) ( 9 )

Related Articles | Metrics

To solve the problem of high characteristic dimension of nonlinear data generated by industrial process, a process fault detection algorithm based on Kernel Principal Component Analysis （KPCA） and Sparrow Search Algorithm （SSA） which is used to optimize the parameters of Support Vector Machine is proposed. Firstly, KPCA algorithm is used to extract linear and nonlinear features of industrial data. Secondly, the data after feature extraction is used as training samples to establish a classification SVM model, and SSA algorithm is used to optimize the kernel parameter and penalty factor of SVM. Finally, the optimized SVM model is applied to test samples for fault detection. In this paper, in order to verify the classification effect of the proposed algorithm, KPCA-SSA-SVM is compared with SVM, KPCA-GA-SVM （Genetic Algorithm, GA） by using a set of nonlinear numerical examples and Tennessee Eastman chemical process data, and the efficiency and superiority of the proposed algorithm is verified.

An Early Diagnosis Method of COVID-19 Infection Based on ResNeXt and Improved nnU-Net

XU Hao, TIAN Zhen-yu, LI Chao-fan, CUI Xin-xin, YANG Jian-lan

2023, 0(06): 21-26. doi:10.3969/j.issn.1006-2475.2023.06.004

Asbtract ( 148 )

PDF (3376KB) ( 40 )

References | Related Articles | Metrics

The early infection of novel coronavirus pneumonia is characterized by increased lung turbidity and density. In order to solve the problem of difficulty in diagnosing and locating lung lesions in early patients with computed tomography, an experimental protocol for the diagnosis of COVID-19 （Corona Virus Disease 2019） with lung lesion segmentation by ResNeXt and a modified nnU-Net （no-new-Net） is proposed. The mean accuracy of ResNeXt model classification is 0.8554, the AUC area is 0.8951, the Precision is 0.8321, the F1 score is 0.8132, and the mean Dice coefficient of improved nnU-Net model lesion segmentation reaches 0.7663, which is a combined improvement of 16.4% compared with other models segmentation ability. The experimental results show that this scheme can enhance the ability to extract infection features from the early lung CT images of new crowns, and achieve efficient disease typing and accurate lesion segmentation.

Modeling Approach of Multi-level and Multi-resolution Grid Model for Strategy Campaign Wargame

LI Hai-yan, WU Da-yu, LIU Qiang

2023, 0(06): 27-32. doi:10.3969/j.issn.1006-2475.2023.06.005

Asbtract ( 218 )

PDF (1273KB) ( 45 )

References | Related Articles | Metrics

The grid model is the core component of strategy campaign wargame environment model and influences other behavior models. Building multi-level and multi-resolution grid models is the basis of extending strategy campaign wargame, supporting fine-grained environment models and key tactical operations. In this paper, the modeling approach and models of multi-level and multi-resolution equal longitude and latitude division for strategy campaign wargame are put forward based on the military requirements and technical requirements. Firstly, grid map is divided by using the equal latitude and longitude quadrangle. Then the coordinate system cluster of equal latitude and longitude grid and related elements are defined. Finally, an example is given. By comparing with single-resolution hexagon grid modeling approach, the advantages are analyzed. The modeling approach can support multiple-resolution of strategy campaign wargame, and meet the requirements of running efficiency and environment model for strategy campaign wargame.

An Experience Replay Strategy Based on Mixed Samples

LAI Jian-bin, FENG Gang

2023, 0(06): 33-38. doi:10.3969/j.issn.1006-2475.2023.06.006

Asbtract ( 123 )

PDF (3744KB) ( 34 )

References | Related Articles | Metrics

Experience replay strategy has become an important part of deep reinforcement learning algorithm. It can not only accelerate the convergence of deep reinforcement learning algorithm, but also enhance the performance of agents. Mainstream experience replay strategies use uniform sampling, priority experience replay, expert experience replay and other methods to accelerate learning. In order to further improve the utilization of experience samples in deep reinforcement learning, this paper proposes an experience replay strategy based on mixed samples （ER-MS）. This strategy mainly uses two methods： immediate learning of the latest experience and review of successful experience. It immediately learns the latest samples generated by the interaction between the agent and the environment, and uses an additional experience buffer pool to save the samples of successful rounds for experience replay. Experiments show that the experience replay strategy based on mixed samples combined with DDPG algorithm can achieve better results in Open AI mujoco task.

Opinion Leaders Mining Method Based on Improved Hits Algorithm

WANG Liu, ZHU Yi-xin, HAN Li-ying

2023, 0(06): 39-42. doi:10.3969/j.issn.1006-2475.2023.06.007

Asbtract ( 153 )

PDF (1065KB) ( 38 )

References | Related Articles | Metrics

Microblog has gradually become an important carrier of public opinion communication. The opinion leaders in online public opinion play a driving role in the process of public opinion communication. It is necessary to explore the opinion leaders in microblog for the management of social network public opinion. Considering microblog users' behaviors such as forwarding comments in the network, a two-layer network of microblog users' “forwarding comments”is constructed. By introducing the influence contribution factor and weight factor of users' interaction behavior to mine users' influence, a microblog users' influence evaluation algorithm based on Hits improved algorithm is proposed. The experimental results show that the F-score comprehensive index score of this model is better than PageRank algorithm and Hits algorithm. It can more accurately identify opinion leaders in microblogging community topics, effectively calculate the actual influence of microblogging users, and more accurately and effectively identify opinion leaders in a certain topic of microblogging community, which can provide reference for research on opinion leaders mining in social networks.

A Collaborative Representation-based Method of Discriminant Locality Preserving Projections

LI Shi-qiu

2023, 0(06): 43-47. doi:10.3969/j.issn.1006-2475.2023.06.008

Asbtract ( 91 )

PDF (1391KB) ( 18 )

References | Related Articles | Metrics

Discriminant locality preserving projections （DLPP） is an effective discriminative feature extraction approach based on manifold learning. It is one of the typical representative algorithms of flow pattern dimensionality reduction. It can use the discriminant information to extract the best discriminant features. However, it fails to exploit the collaborative reconstruction relationship between samples, and thus usually leads to a lower recognition rate. To cope with this problem, a collaborative representation based discriminant locality preserving projections （CRDLPP） is proposed. CRDLPP first calculates the collaborative reconstruction errors of all the samples by collaborative representation mechanism, and then incorporates them as a regularization term into the objective function of DLPP. The optimal projection matrix is finally obtained by solving the new objective optimization problem. In order to verify the good performance of CRDLPP method in image recognition, this paper selects public image databases such as Yale and coil20 for experiments. The results show that the CRDLPP algorithm in this paper has a higher recognition rate than other classical data dimensionality reduction algorithms in image recognition.

Domain Adapted Person Re-identification Algorithm Based on Joint Network

LI Guo-xin, QU Han-bing, ZHU Cheng-bo, WANG Xin-xuan, HU Jia-bao

2023, 0(06): 48-55. doi:10.3969/j.issn.1006-2475.2023.06.009

Asbtract ( 88 )

PDF (4535KB) ( 31 )

References | Related Articles | Metrics

Aiming at the problems of insufficient use of image details in the domain adaptation method and low accuracy of cross-domain person re-recognition due to data distribution differences between domains, a domain-adapted person re-recognition algorithm based on joint network was designed. Firstly, in order to improve the model's ability to obtain details, an attention mechanism module is embedded in the network, and an asymmetric multi-granularity network was designed to supplement the missing details in the global feature. Secondly, in order to alleviate the influence caused by inter-domain differences, online correlation consistency loss is introduced on the basis of CycleGAN network, and sample generator is trained to generate source domain samples with target domain style to reduce inter-domain data distribution differences. Then, in the domain adaptation stage, the method based on clustering joint network is designed, and the classification probability output by student model in the joint network is used as the supervision information to supervise the training of teacher network, so as to avoid the error amplification caused by clustering hard pseudo-label in training. Finally, Momentum Contrast Loss （MoCo Loss） is introduced to alleviate the influence of noise pseudo-labels to improve the adaptive ability of the model. Experiments are carried out on Market-1501 and DukeMTMC-Reid, and the results show that the algorithm has certain competitiveness compared with the mainstream algorithms.

Alzheimer’s Disease Image Classification Based on Improved EfficientNet

ZHU Jian-bo, GE Ming-feng, DONG Wen-fei

2023, 0(06): 56-61. doi:10.3969/j.issn.1006-2475.2023.06.010

Asbtract ( 149 )

PDF (2105KB) ( 51 )

References | Related Articles | Metrics

To improve the effectiveness of the convolutional neural network for Alzheimer’s disease MRI image classification, a convolutional neural network FAMENET is proposed, which integrates an adaptive attention mechanism and data enhancement technique to alleviate data imbalance by introducing a data augmentation technique and Focal Loss loss function. The network is reconfigured to reduce the number of model parameters and the computational effort of the network while maintaining accuracy. The adaptive attention mechanism is introduced to solve the information loss problem caused by the downsampling of input images for feature extraction. In a large number of comparative experiments on public datasets, the classification accuracy of FAMENET reaches 79.95% and the AUC value reaches 82.54%. The designed ablation experiments also fully demonstrate the effectiveness of the proposed modules and networks.

An Improved Algorithm of Removing Adherent Noise Based on Binocular Vision

PENG Ming-kang, FENG Cheng-de

2023, 0(06): 62-68. doi:10.3969/j.issn.1006-2475.2023.06.011

Asbtract ( 128 )

PDF (1626KB) ( 31 )

References | Related Articles | Metrics

In order to remove the image noise caused by the attachments on the lens or protective glass, a fast denoising algorithm based on binocular vision is proposed. Firstly, by introducing the step-by-step matching and integral map, the fast calculation of the real disparity map is realized. Secondly, the theoretical disparity of the attached noise is calculated by using the distance between the binocular camera and the glass where the attachment is located, and the noise localization is completed in combination with the actual disparity map. Then, image inpainting is used to estimate the disparities of the noise area, and the disparities of the pixels occluded by the attached noise is obtained. Finally, for the noise pixels in the left image, the estimated disparity is used to find the corresponding clean pixels in the right image and replace them to complete the removal of attached noise. The experiment compares the algorithm with the existing binocular vision-based denoising algorithm, and the results show that the algorithm can significantly improve the denoising speed while ensuring similar denoising effect. Compared with the learning-based denoising algorithm, this method has the advantages of stable denoising effect and will not identify other areas as noise.

Road Pothole Detection Algorithm Based on Improved YOLOv5s

BAI Rui, XU Yang, WANG Bin, ZHANG Wen-wen

2023, 0(06): 69-75. doi:10.3969/j.issn.1006-2475.2023.06.012

Asbtract ( 551 )

PDF (3457KB) ( 58 )

References | Related Articles | Metrics

Aiming at the problem that existing target detection algorithms are difficult to accurately detect road potholes and the detection speed is slow, a road pothole detection algorithm based on improved YOLOv5s is proposed. Firstly, CA （Coordinate attention） module is integrated into YOLOv5s backbone network, so that the model can capture not only cross-channel information, but also direction perception and position sensitive information, which is helpful for the model to locate and identify the detected object more accurately. Then, SoftPool is adopted in Spatial Pyramid Pool （SPP） module to improve the maximum pooling operation and retain more detailed characteristic information. In the feature fusion stage, Content-Aware ReAssembly of FEatures （CARAFE） is used to improve the up-sampling of multi-scale feature fusion and dynamically generate an adaptive kernel, which can gather context information in a large receptive field. Finally, Alpha-IoU is used to improve the loss function and improve the margin regression accuracy. Experimental results show that the average accuracy of the improved YOLOv5s algorithm is 4.6 percentage points higher than that of the original network, and the detection accuracy of the improved YOLOv5s algorithm is greatly improved compared with other mainstream algorithms such as SSD, Faster R-CNN, YOLOv3, YOLOv3-tiny and YOLOv4-tiny.

Monocular Depth Estimation Method by Aggregating Multi-dimensional Attention Features

LIU Jia-jia, HU Xu-xin, YU Ping

2023, 0(06): 76-81. doi:10.3969/j.issn.1006-2475.2023.06.013

Asbtract ( 88 )

PDF (1967KB) ( 51 )

References | Related Articles | Metrics

This study is outlined to improve the precision for predicting monocular depth estimation networks and provides an in-depth analysis of the effects of multidimensional attention mechanisms on monocular depth estimation networks. The conclusions and observations are used to design a set of optimized channel and space attention blocks. Considering the convolutional neural network framework obtained based on the local plane guidance layer, a new network structure is created to fully activate the multidimensional attention mechanism through a method that is based on placing different design blocks. Furthermore, in combination with the above two measures for improvement, this study proposes a high-performance monocular depth estimation network that integrates channel and space attention features. On the KITTI Depth dataset and an NYU Depth V2 dataset, the outcomes of this study prove the effectiveness of the optimized blocks and the satisfactory performance of the proposed network through experiments. Compared with the convolutional neural network based on the local plane guidance layer, the proposed network is better in processing the overall features of images and more accurate in predicting depth information with several metrics for network evaluation improved to different degrees. The depth maps generated by the proposed network also demonstrated more data associated with the contours and details of objects.

A Cascaded Insulator Defect Detection Model Combining Semantic Segmentation and Object Detection

YE Li-ming, CHEN Wei-wen

2023, 0(06): 82-88. doi:10.3969/j.issn.1006-2475.2023.06.014

Asbtract ( 114 )

PDF (1983KB) ( 59 )

References | Related Articles | Metrics

Insulator defect detection is an important part of the routine inspection of substations. Using the video surveillance images in the substation, we propose a cascaded insulator defect detection model that combines semantic segmentation with object detection aiming at the low detection accuracy of the insulator defect detection model. The model consists of three modules： Insulator segmentation, insulator cutting and defect detection. The insulator segmentation module separates the insulator from the complex environment and proposes an edge enhancement loss function. The insulator cutting module uses the image processing method to obtain the insulator region aligned with the axis. The defect detection module completes the defect detection. Experiment results show that the accuracy of edge-enhanced Unet for insulator segmentation reaches 83.99%, and the accuracy of RetinaNet with improved anchor generation method for defect detection reaches 63.32%. Compared with the single-stage insulator defect detection model, the proposed cascade insulator defect detection model can effectively eliminate the interference of environment, detecting most of the insulator defects.

Medical Insurance Fee Control System Based on Data Mining

LIU Pei

2023, 0(06): 89-94. doi:10.3969/j.issn.1006-2475.2023.06.015

Asbtract ( 134 )

PDF (2002KB) ( 51 )

References | Related Articles | Metrics

Medical insurance fee control can ensure the stable expenditure of national medical insurance fund, but the function and performance of medical insurance fee control system can not meet the requirements of users. This paper designs a medical insurance fee control system based on data mining. Using the RF transmitter, the medical insurance expense data synchronizer is designed. Combined with the medical insurance expense data extraction circuit and medical insurance expense controller, the hardware design of the system is completed. In the software design, according to the attributes of medical insurance expense data, the gain value of medical insurance expense information is calculated, and the characteristics of medical insurance expense data are extracted by using data mining algorithm. Through the classification of outpatient medical insurance expenses, the medical insurance expense control algorithm is designed to realize the control of medical insurance expenses. The test results show that the medical insurance expense control function of the system meets the requirements of users, and can control the proportion of illegal amount and memory occupancy within 17% and 20% respectively, which greatly improves the performance of the system.

Hybrid Cloud Data Architecture for Civil Aircraft Manufacturing

LU Wei-qiang

2023, 0(06): 95-102. doi:10.3969/j.issn.1006-2475.2023.06.016

Asbtract ( 101 )

PDF (11300KB) ( 26 )

References | Related Articles | Metrics

The transformation of digital technology has continuously pushed the traditional manufacturing industry to become digital and intelligent. As a high-end manufacturing industry, the civil aircraft industry has a long development cycle, complex business, and extremely high requirements for supplier collaboration. The digital transformation of the civil aircraft manufacturing industry is a historical opportunity to lead the country's manufacturing industry to the forefront of the world. Data is the core of digital transformation and the core asset of an enterprise. Taking an aircraft manufacturing enterprise as an example, according to the characteristics of civil aircraft manufacturing industry, the design framework of hybrid cloud data architecture is given. The business model is obtained from the domain modeling led by business requirements, and the data model and microservices are independently developed based on this, and the hybrid cloud data architecture is designed to form a complete closed loop with clear requirements, reasonable business split, and light and fast service development and deployment. Under this data structure, 13 types of problems and 12 core subsystems were initially managed for data, and 6 databases, 9900 table objects and other data assets were obtained, and the governance effect was obvious.

Efficient Healing Key Management Scheme for UAV Group

LIU Jun, YUAN Lin, FENG Zhi-shang, ZHANG Biao, LIU Chao

2023, 0(06): 103-109. doi:10.3969/j.issn.1006-2475.2023.06.017

Asbtract ( 128 )

PDF (1878KB) ( 35 )

References | Related Articles | Metrics

Unmanned Aerial Vehicle Group Network （UAVGN） has dynamic topology and high mobility, so it is easy to suffer security threats and attacks in the open air wireless environment. Especially the communication transmission link is prone to interference, and the communication is unstable or even is interrupted, leading to the loss of key packets in the process of key management, so that the subsequent session cannot establish key secure communication. Therefore, considering the limited UAV resources, this paper proposes a healing group key management scheme based on Hash chain and Chinese residual theorem, which has two mechanisms of self-healing and mutual healing, and improves the flexibility and efficiency of UAV key updating. Meanwhile, group key update is divided into local update key and session update key. The local update key is updated locally by the preset Hash function, which further reduces the calculation overhead of key update and improves the key processing capability of UAV. The Key Group Manager （KGM） selects the Key update factor in the session stage, constructs the broadcast message of key update by using the Chinese Remainder Theorem （CRT）, and realizes the dynamic and flexible update of key. The analysis and experiment show that the scheme has security performance such as forward security, backward security, anti-undo ability, anti-collusion attack and anti-replay attack. Compared with the existing schemes under the same conditions, this paper effectively optimizes the computation and communication overhead, improves the key update efficiency, and can solve the key update problem in the unstable communication of UAVGN, thus ensuring the establishment of secure key communication.

Survey on Blockchain Security Protection

ZHA Kai-jin, WANG Zhi-bo, HE Yue-shun, XU Hong-zhen

2023, 0(06): 110-117. doi:10.3969/j.issn.1006-2475.2023.06.018

Asbtract ( 228 )

PDF (1274KB) ( 62 )

References | Related Articles | Metrics

Blockchain technology, as one of the most popular technologies at present, has huge application value. At the same time, it is widely used in many key fields due to the high support of the country. The many characteristics of blockchain technology determine its application advantages in data sharing, digital storage, information tracing, security guarantee, etc., and at the same time, it also brings many security risks. Because of this, this article summarizes the content and related research progress of the blockchain infrastructure, security threats, and privacy protection schemes by studying on high-quality literature on blockchain security protection related research at home and abroad. Aiming to the development status of blockchain privacy protection technology, from the two aspects of encryption technology improvement and privacy protection technology fusion research,we analyze its impact on the development of blockchain, hoping to provide reference for blockchain security protection research.

Commercial Cryptographic Upgrade for Industrial Internet Platform

MO Yan, TANG Rong-chuan, JU Hao, SUN Shao-fei, WANG An

2023, 0(06): 118-126. doi:10.3969/j.issn.1006-2475.2023.06.019

Asbtract ( 146 )

PDF (6846KB) ( 35 )

References | Related Articles | Metrics

Industrial Internet is an important part of key information infrastructure, and its information security is very important. Commercial cryptographic algorithm is an important means to ensure the information security of industrial Internet. In response to these two types of security problems, this paper proposes a basic architecture for applying commercial cryptography to the industrial internet platform, and provides a solution for upgrading and transforming cryptographic applications at physical layer, infrastructure-as-a-service, platform-as-a-service, software-as-a-service, device layer and control layer of the industrial internet platform. It has been widely used to solve various security problems in information security management and information security technology, and the architecture has been applied to the implementation process of actual projects in the industry. The cryptographic application upgrade and transformation framework is connected and reconstructed with the existing business framework of the Hanyun Industrial Internet Big Data Platform and the Xuzhou Construction Machinery Group’s （XCMG） Internet of Vehicles Platform. The compatibility of the original platform equipment and the commercial cryptographic products that meet the requirements of relevant national standards has been achieved. The upgrade and transformation of the cryptographic application of the Hanyun Industrial Internet Big Data Platform and the XCMG Internet of Vehicles Platform has been achieved which demonstrated the broad applicability and sustainable development potential of the architecture.

Table of Content