最新刊期

    22 11 2023
    • TAO Liufeng,XU Yongyang,WU Yehui,CHEN Zhanlong,WU Liang,XIE Zhong
      Vol. 22, Issue 11, Pages: 1-8(2023) DOI: 10.11907/rjdk.231961
      摘要:With the continuous expansion of the application fields and models of earth information systems, the related technologies and software of earth information systems are increasingly developing. However, there are still issues such as the efficiency of constructing full space three-dimensional models, the integration of multiple data types and complex relationships, and the rendering of large-scale multi-source heterogeneous data. Therefore, based on the MapGIS platform, the key technologies of the full space earth information system are mainly elaborated from four aspects: cognitive modeling and visualization of full space GIS, big data management, intelligent analysis, and intelligent mapping. Abstract real-world entities from the perspectives of multiple granularity, types, changes, and correlations, and model them to form a cognitive model of the entire space. Starting from the characteristics of massive, diverse, and real-time, this article elaborates on the geographic big data management technology required for full spatial GIS. Elaborate on full space intelligent analysis technology from the perspectives of knowledge driven, artificial intelligence, and data science computing. From the perspective of intelligence, this article elaborates on the full space map mapping technology, MapGIS platform architecture, and its application and promotion value, providing reference for the development of the full space earth information system platform technology.  
      关键词:pan-spatial;GIS;spatial-temporal big data;geographic knowledge graph;modeling and visualization   
      116
      |
      32
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545263 false
      发布时间:2023-11-27
    • ZOU Xiuzhen,HU Yuchen,ZHU Changqing
      Vol. 22, Issue 11, Pages: 9-17(2023) DOI: 10.11907/rjdk.231737
      摘要:As the key data of national infrastructure, the importance of copyright protection of geographic information is self-evident. However, the existing digital watermarking systems for geographic information are all developed based on Windows operating system, which cannot realize the independent control of the software and meet the demand of copyright protection of geographic information on Linux operating system. Therefore, based on the open-source Geospatial Data Abstraction Library (GDAL), this paper develops a geographic information digital watermarking system for Linux operating system by using Java language and Swing framework, and introduces the system and its implementation details from the aspects of system design, key technologies, system implementation, etc. . The system consists of four parts: watermark embedding, watermark detection, log management, and permission management, and it designs digital watermarking algorithms for different geographic information data formats and solves the problem of reading and writing data in ESRI FileGeodataBase (FileGDB) in the process of implementation. The experimental results show that the system is imperceptible and the algorithm is robust against common attacks. The implementation of this system makes up for the blank of the current geographic information copyright protection software in Linux system, truly realizes the autonomy and control of GIS software, and makes a positive contribution to promoting the localization of GIS software.  
      关键词:copyright protection;geographic information;digital watermarking;software design;Linux operating systems   
      86
      |
      14
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545265 false
      发布时间:2023-11-27
    • CHEN Genshen,LIU Gang,CHEN Qiyu,ZHAO Lixing,XIANG Shize,LUO Qing,ZHANG Junjie
      Vol. 22, Issue 11, Pages: 18-28(2023) DOI: 10.11907/rjdk.231777
      摘要:Mining production processes, comprising different fields such as geology, geophysics, drilling, exploration, and geomatics, confronts with significant challenges in terms of spatial data integration and management complexities. Therefore, this paper introduces a 3D spatial mining management system based on multi-source data integration and WebGIS, in which digital information derived from geology, geophysics and drilling is mainly used to support establishment of unified comprehensive geological information database so that restoration, transformation, management, querying, analysis, and visualization of spatial data, attribute information, model objects, and temporal data. In practical application, according to integrating 3D reality models, geological models, seismic data models and multiple systems data fusion to form a comprehensive mining production framework. This framework efficiently manages and share geological information during mineral exploration, thereby providing robust technological support for mining production decision-making.  
      关键词:multi-source data integration;digital mine;WebGIS;three-dimensional visualization   
      52
      |
      6
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545053 false
      发布时间:2023-11-27
    • WEI Rulan,WANG Hongfei,SHENG Sen,JIANG Yifan,YU Yafang
      Vol. 22, Issue 11, Pages: 29-34(2023) DOI: 10.11907/rjdk.231908
      摘要:In order to stick to the red line of 1.8 billion mu (the Chinese version of acre) of cultivated land, China has introduced a series of policies and regulations to resolutely implement the strictest cultivated land protection system. How to quickly discover the change of cultivated land is a key step of the cultivated land protection system. The AI based change detection method for satellite remote sensing images is mainly based on multi temporal change detection. Currently, there have been many breakthroughs in theory, but there are still some difficulties in practical application. Hence, this paper proposes an easy-to-use method to discover the change of cultivated land based on the high-resolution satellite images and the annual change survey results: Firstly, a semantic segmentation model(RABD)fusing multi-scale features is used to quickly extract the cultivated land patches;After post-processing and grid erasing processing with the previous farmland vector pattern, a suspected farmland reduction pattern was obtained. And based on this method, it is packaged into software, then, the cultivated land patches after post-processing is grid-based erased with previous cultivated land vector patches to obtain the suspected cultivated land reduction patches. And based on this method, packaged into software, it can quickly detect the cultivated land patches occupied by construction, shorten the manual operation time by about 70%, and provide an intelligent technology choice for cultivated land protection.  
      关键词:cultivated land change detection;illegal occupation;deep learning;semantic segmentation;high-resolution image   
      79
      |
      5
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544975 false
      发布时间:2023-11-27
    • JIA Ding,CHEN Xiaohong,ZHOU Yonggui
      Vol. 22, Issue 11, Pages: 35-41(2023) DOI: 10.11907/rjdk.231778
      摘要:GNSS (Global Navigation Satellite System) is widely used in the field of geological disaster monitoring and warning. The improved tangent angle calculated from its displacement time curve is an important criterion in landslide monitoring and warning models. The existing improved tangent angle calculation methods have high fluctuation in the calculation results, and the selection of velocity during the uniform deformation stage has strong randomness, resulting in low calculation accuracy and timeliness. This article implements an improved tangent angle algorithm, which uses the five point method of Lagrange interpolation differentiation to improve computational accuracy, increasing by 77.33%. And through the uniform speed stage identification method and setting speed update conditions, the dynamic real-time calculation of tangent angle is achieved, providing real-time and dynamic criteria for landslide monitoring and warning, and improving the practicality of landslide monitoring and warning.  
      关键词:GNSS;tangent angle;displacement-time curve;landslide;monitoring and early warning   
      16
      |
      2
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544893 false
      发布时间:2023-11-27
    • CHEN Songbin,HU Cong,WANG Junyong,CHENG Yuyang
      Vol. 22, Issue 11, Pages: 42-48(2023) DOI: 10.11907/rjdk.231731
      摘要:The sharing of government information resources has played an important role in reducing the cost of repeated information collection and improving the efficiency of information sharing. However, currently, the sharing of government information resources is facing issues of low coverage and inefficiency in data integration caused by cross domain entity mutual recognition. To address this issue, a government information resource integration model based on urban digital identification is proposed, aiming to build a new framework for cross domain data fusion and sharing. By utilizing urban digital identification technology to uniquely identify and manage elements such as people, events, and objects within the city, the uniqueness, security, and sharability of cross domain data identification are achieved through complementary advantages of "code+chain". By analyzing two common modes of government information resource integration, this paper proposes the design and implementation process of a one number one source, source governance, and data integration driven business integration mode based on urban digital identification. This model integrates key elements such as people, housing, enterprises, time, and space, and supports government applications in multiple fields such as market regulation, social management, and public services through data aggregation and integration, sharing and openness, and development and utilization. The practical application case has verified that this model can effectively integrate government information resources, thereby improving the collaboration efficiency between departments.  
      关键词:e-government;government big data;geospatial information;data fusion;data sharing;social governance   
      8
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545261 false
      发布时间:2023-11-27
    • MING Zi,LIU Wei,LI Yang,CUI Junjie,LIU Gang,LI Jiahui,LEI Mengting
      Vol. 22, Issue 11, Pages: 49-56(2023) DOI: 10.11907/rjdk.231773
      摘要:Regional query technology is an important component of urban geospatial big data GIS systems. When facing multi center and non-uniformly distributed urban geographic spatial data, compared to the traditional R-tree index with a tree structure, MVD index effectively avoids the low efficiency of spatial query algorithms caused by node overlap through a multi-layer network structure. However, native MVD index cannot directly support regional queries of non point spatial objects. To address this issue, firstly, a non point spatial data management scheme based on MVD index is designed based on the mapping relationship between point centroid object as a whole. Based on this scheme, a reliable spatial region query algorithm MVD-Polygon is implemented using the boundary extension technology of the minimum outer circle of spatial objects; Then, based on the scale distribution characteristics of spatial objects, a hierarchical index construction strategy is proposed to improve the efficiency of regional query algorithms; Finally, based on the MVD index, a high-performance regional query technology framework MVD-Polygon-Grade was formed for urban geospatial non point data. The experimental results show that compared with the mainstream region query algorithm Multi-step based on R-tree index, MVD-Polygon-Grade has higher operational efficiency in urban geographic spatial query scenarios, fully proving the feasibility and practicality of the proposed technical framework.  
      关键词:non-point spatial object;MVD index;region query   
      17
      |
      3
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545123 false
      发布时间:2023-11-27
    • QI Xiangming,YE Songfa
      Vol. 22, Issue 11, Pages: 57-64(2023) DOI: 10.11907/rjdk.231584
      摘要:The real-time detection of insulator string dropping plays an important role in ensuring the stability of high-voltage power transmission and timely troubleshooting. Therefore, an aerial insulator string dropping detection method was proposed based on lightweight YOLOv5-GCDNet. Firstly, a new feature extraction network, GCDNet, was designed using ghost convolution and inverted residual structure to reduce the computational complexity of YOLOv5 target detection network, improve the detection speed, and extract more abundant target features. Secondly, the inverted residual structure was used to improve the feature extraction network to extract richer target features, and the adaptive upsampling was combined to optimize the feature fusion network to improve the model detection accuracy. Finally, the α-IoU excitation factor was added to accelerate the convergence of the target frame loss value to improve the detection efficiency. Through the comparison of the experimental results, the average detection accuracy of the lightweight YOLOv5-GCDNet model proposed in this paper was 93.6%, the detection speed was 68FPS, and the parameters of the model was 3.7M, which was better than other detection algorithms under the same conditions. The method in this paper can provide a technical reference for the insulator string dropping detection.  
      关键词:string-drop insulator;defect detection;YOLOv5;ghost convolution;inverted residual;adaptive upsampling   
      32
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544398 false
      发布时间:2023-11-27
    • DONG Chen,ZHENG Lu,YU Shu,RAO Baiyun
      Vol. 22, Issue 11, Pages: 65-70(2023) DOI: 10.11907/rjdk.231470
      摘要:Remote sensing images have multi-scale characteristics due to the presence of objects of varying sizes. Most existing change detection methods use multi-scale segmentation to obtain more complete and rich local features, but ignore the importance of non local position features (global context information), resulting in weak differential representation of extracted features. To address this issue, a twin network SPAN based on context awareness and multi-scale attention is proposed. In the feature extraction module, twin pyramid structures are used to extract features of different scales for fusion to obtain more robust features; In the attention enhancement module, an attention module PSG that combines global context pyramid segmentation is introduced to further extract target features and enhance differential representation. The experiment shows that compared to existing models, the F1 of the proposed model on the change detection standard dataset CDD and LEVIR-CD reaches 0.944 and 0.812, respectively. The ablation experiment shows that PSG can fully extract discriminative features compared to traditional multi-scale attention, demonstrating the importance of global contextual information for change detection tasks.  
      关键词:change detection;siamese;multi-scale;attention;global context   
      17
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544462 false
      发布时间:2023-11-27
    • LIAO Xuanzhi,QIU Qian,XIONG Xianming
      Vol. 22, Issue 11, Pages: 71-78(2023) DOI: 10.11907/rjdk.231756
      摘要:Due to the current lack of specialized software for satellite remote sensing image analysis of shallow water depth, this article studies a shallow sea remote sensing water depth inversion software based on IDL, so as to invert shallow water depth quickly and accurately. Firstly, the theoretical research on satellite remote sensing water depth inversion at home and abroad was analyzed, as well as the current status of professional software for water depth inversion, on this basis, relying on IDL language, designed and implemented a shallow sea remote sensing water depth inversion software that includes input display, water depth inversion, output saving, and other auxiliary modules. Absolute water depth and relative water depth were retrieved using IKONOS at the harbor near Nanliao, Hsinchu City and GeoEye-1 Satellite imagery at Hongxiu Island, which belongs Nansha Islands. Various evaluation indicators under absolute and relative water depth inversion, Among them, R2 is 0.996 7, 0.989 7, RMSE is 0.572 6 m, 1.076 1 m, MRE is 6.659 3%, 5.496 2%, MAE is 0.479 2 m, 0.659 2 m. It can be seen that the accuracy of water depth inversion is good, and the software has high stability, achieved fast and accurate shallow sea remote sensing water depth inversion. The software has a certain degree of popularization.  
      关键词:inversion software;water depth inversion;shallow sea remote sensing;IDL   
      12
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544705 false
      发布时间:2023-11-27
    • WU Liang,SONG Zhenzhen,CHEN Zhanlong,ZHOU Qi,GUO Mingqiang,LIU Zheng
      Vol. 22, Issue 11, Pages: 79-84(2023) DOI: 10.11907/rjdk.231614
      摘要:This study aims to investigate and implement talent cultivation modes for geospatial information software in the era of intelligence. The primary objective is to enhance students' comprehensive abilities to address increasingly complex real-world problems and meet the requirements of modern geospatial information software development in terms of talent capabilities and qualities. To achieve this, an integrated interdisciplinary teaching platform is constructed, combining artificial intelligence with the field of geospatial information engineering. A new curriculum structure is established, and an innovative teaching model is developed, emphasizing intelligence, multidimensionality, and practicality. The interdisciplinary cultivation model is employed to foster talent with cross-disciplinary expertise. Disciplinary development and talent cultivation mutually support each other, and exploring talent cultivation models for geospatial information software in the era of intelligence presents more opportunities than challenges. It is crucial to clarify the objectives of disciplinary development, optimize talent cultivation programs, and enrich curriculum and practical resources. This ensures that the outcomes of disciplinary practice meet the demands of educational reforms, the development of an enriching academic atmosphere, innovation practices, and the training of exceptional engineers in the field of geospatial information software.  
      关键词:era of intelligence;geospatial information software;disciplinary development;talent cultivation model;teaching innovation mode   
      16
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544895 false
      发布时间:2023-11-27
    • ZHANG Yinsheng,SHAN Hinlin,HE Bingkun,GUO Zimin
      Vol. 22, Issue 11, Pages: 85-90(2023) DOI: 10.11907/rjdk.231133
      摘要:Aiming at the requirements of machine learning courses and combining with practical engineering problems of electronic information object detection, a deep learning based object sorting teaching experimental system is proposed. Identify and classify the targets captured by the Jetson Nano camera, use a control chip to convert the current position of the targets into three-dimensional coordinates, and control the robotic arm to grasp and sort them. At the same time, YOLOv3 target detection algorithm is improved, MobileNetV2 is used to replace the feature extraction network to reduce the feature scale, the adjusted Focal loss loss function replaces YOLOv3 loss function, and the anchor box clustering algorithm is replaced by K-Means++. The experiment shows that the average accuracy of the improved algorithm is increased by 23.4%, the average detection time is reduced by 9.35 ms, and the detection accuracy and speed are significantly improved. It can meet the practical teaching requirements of advanced technologies such as deep learning and machine vision, and help to enhance students' innovation and practical abilities.  
      关键词:experiment system;deep learning;target sorting;mechanical arm;Jetson Nano   
      19
      |
      2
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544252 false
      发布时间:2023-11-27
    • TAN Xiaohui,QI Zhenghua,ZHANG Haitao,HE Feifei
      Vol. 22, Issue 11, Pages: 91-97(2023) DOI: 10.11907/rjdk.222448
      摘要:Aiming at the problem that the existing neighborhood-based session recommendation methods ignore the general preference, current preference, and anonymous user misclick behavior of the whole session. A session recommendation method based on neighborhood and position time awareness is proposed. The proposed recommendation method considers three factors, namely the duration and location of items in the current session, temporal correlation between sessions, and the lasting time and position of candidate items in neighboring sessions. Experimenting with the recommendation method with the benchmark method on two public news datasets, Globo.com and Adressa, the experimental results show that the recommendation method improves on both HR@20 and NDCG@20 metrics compared to the best-performing benchmark method. Especially on the Globo.com dataset, the HR@20 and NDCG@20 indicators were improved by 4.3% and 7.0%, respectively, indicating that the proposed method can effectively improve the recommendation effect.  
      关键词:session recommendation;position-aware recommendations;time correlation;misclick behavior   
      25
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544548 false
      发布时间:2023-11-27
    • ZHENG Xinyi,ZHANG Yushan
      Vol. 22, Issue 11, Pages: 98-103(2023) DOI: 10.11907/rjdk.231634
      摘要:In order to solve the problem that the existing recommendation models often give priority to the accuracy of recommendation and ignore the diversity and novelty of recommendation, which leads to the fact that the recommendation results are always single and inevitably occupied by popular items, this paper proposes a reordering model. Based on the original recommendation model, this model adds item popularity and user preference, and controls the proportion of novel items in the final TOP-K recommendation list by adjusting the compensation score, thus improving the diversity of personalized recommendation list. The experimental results on Movielens-1M data set and FilmTrust data set show that the ILS index of this model is reduced by 79.96% and 7% respectively, and the Novelty index is reduced by 0.73% and 2.11% respectively, which is helpful to reduce the similarity of users in the list and increase the novelty of recommendation.  
      关键词:recommender systems;diversity;re-ranking;item popularity;recommendation novelty   
      29
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544809 false
      发布时间:2023-11-27
    • YU Xiaoxue,WEI Jianwen,ZHANG Zhanbing,GAO Yiqin,LIN Xinhua,ZHOU Yun,ZHOU Yanxiao
      Vol. 22, Issue 11, Pages: 104-109(2023) DOI: 10.11907/rjdk.231905
      摘要:Machine learning approaches are widely adopted in job runtime prediction tasks. However, how to collect data, the choice of ML methods, tricks to use in training, plus how to improve accuracy on small data sets are obstacles for real use. This paper implements a SLURM-based job collection system for Gaussian, one of the most popular scientific applications on computational learning platform. Multiple ML and feature extraction methods are evaluated. Experiments show that, on a data set containing 5 314 samples, the methods based on decision tree and deep neural network combined with feature extracted by Text-CNN and F-norm of coulomb matrix obtained 71.87% and 84.93% accuracy respectively, surpassing traditional linear regression and support vector machine methods. On two small data sets with only 118 samples, the transfer migration learning method which reuses the pre-training parameters, achieves 65.2% and 63.3% accuracy, surpassing that of training from. Using Text-CNN can avoid sparse matrix caused by one-hot encoding and achieve a higher scalability. Besides, the F-norm of the Coulomb matrix played an important role in improving the accuracy.  
      关键词:machine learning;computational learning platform;job scheduling systems;job runtime;Gaussian application   
      13
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544891 false
      发布时间:2023-11-27
    • ZHANG Wenjie,GAO Guoqin
      Vol. 22, Issue 11, Pages: 110-117(2023) DOI: 10.11907/rjdk.222431
      摘要:In order to solve the problem of jet reaction force mismatching disturbance at the end of the spray gun of the parallel mechanism for sandblasting and derusting, an adaptive backstepping sliding mode control method for the parallel mechanism with expanded state observer is proposed. Firstly, by using the bar length constraint and geometric relationship of the parallel mechanism, the kinematics of Stewart 6-DOF parallel mechanism is analyzed by analytical method, and the dynamic model of parallel mechanism with unmatched interference sandblasting derusting with jet reaction force is established by Lagrangian method. Then an extended state observer is designed to realize the on-line estimation of the mismatched disturbance, and the intermediate virtual control quantity is designed based on the estimated value of the mismatched disturbance in the backstepping design process to compensate the mismatched interference of the parallel mechanism system. the switching gain of sliding mode control is dynamically adjusted by designing adaptive law to achieve the effect of suppressing chattering of sliding mode control. The simulation and experimental results show that this method can make the sandblasting derusting parallel mechanism have good system robustness and trajectory tracking effect under the strong interference of sandblasting jet reaction force.  
      关键词:parallel mechanism;unmatched interference;extended state observer;backstepping;sliding mode control   
      12
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544460 false
      发布时间:2023-11-27
    • SUN Xiaotian,ZENG Qingjun,SHANG Le,LIU Zihao,DAI Xiaoqiang
      Vol. 22, Issue 11, Pages: 118-122(2023) DOI: 10.11907/rjdk.222440
      摘要:A thrust allocation based fault-tolerant control method for the thruster of the developed "Sea Exploration Type I" autonomous underwater robot is proposed to address the issue of thruster failures in complex environments. Firstly, establish a geodetic coordinate system and a motion coordinate system to describe the dynamic model of a six degree of freedom autonomous underwater robot in space, and based on this, design a backstepping sliding mode motion tracking controller. Combined with the Lyapunov stability criterion, prove the asymptotic stability of the system; Then, the thrust allocation model is reconstructed based on different types and degrees of faults to compensate for the efficiency of the faulty thruster losses. The simulation results show that this method can make the actual motion trajectory of the AUV tend to the expected trajectory, efficiently achieving accurate motion tracking of the AUV. When the thruster fails, it can compensate for the thrust loss caused by the AUV failure in a short period of time, allowing the AUV to return to the expected path and continue underwater operations.  
      关键词:autonomous underwater vehicle;thrust distribution;backstepping sliding mode;fault-tolerant control   
      12
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544319 false
      发布时间:2023-11-27
    • CHEN Jiwei,BAO Changchun,ZHAO Ziheng
      Vol. 22, Issue 11, Pages: 123-128(2023) DOI: 10.11907/rjdk.222449
      摘要:An improved A * algorithm is proposed to address the problems of large map size, high computational memory consumption, and long computational time in path planning for unmanned aerial vehicle logistics delivery environments. This method includes incorporating a jump point search algorithm search strategy, designing a starting point search method, improving the algorithm heuristic function, and simulating and analyzing the A * algorithm, jump point search algorithm, and improved A * algorithm by constructing maps of different sizes. The results show that the improved A * algorithm maintains the optimal path of A * algorithm, with an average reduction of 90% in search time compared to A * algorithm and 55% in jump point search algorithm, confirming the efficiency and feasibility of the improved A * algorithm.  
      关键词:logistics distribution;path planning;A* algorithm;jump point search algorithm;optimal path   
      12
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544251 false
      发布时间:2023-11-27
    • CHEN Zhigang,ZHENG Shenwen,SHEN Yue
      Vol. 22, Issue 11, Pages: 129-132(2023) DOI: 10.11907/rjdk.222270
      摘要:Aiming at the torque ripple of permanent magnet synchronous motor at low speed, a fuzzy fractional PID iterative learning control strategy is proposed. Firstly, the source of torque ripple of permanent magnet synchronous motor is analyzed and the mathematical model is established. Then, on the basis of iterative learning control, in order to solve the problem of limited adjustment range of the controller, a fractional-order PID type iterative learning control scheme is proposed, which extends the order of calculus from integer to any real number, and introduces fuzzy control theory to solve the problem of difficult parameter selection to realize real-time online adjustment of the controller parameters. The experimental results show that the proposed control algorithm can effectively suppress the output torque ripple, improve the dynamic performance and stability of the control system, and present a stronger anti-interference performance, in order to provide new methods and ideas for solving the torque ripple problem of permanent magnet synchronous motor.  
      关键词:permanent magnet synchronous motor;iterative learning control;fractional order;fuzzy control;torque ripple   
      8
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544247 false
      发布时间:2023-11-27
    • HUANG Yongguo,CHEN Xiaohong,SUN Fujia,LIN Xiaojia,ZHAO Haonan
      Vol. 22, Issue 11, Pages: 133-140(2023) DOI: 10.11907/rjdk.222379
      摘要:During the operation of the ultrasonic scalpel, the resonant frequency of the ultrasonic transducer drifts due to the change of acoustic impedance and self-heating of the ultrasonic scalpel, which makes it difficult for the frequency of the driving signal to accurately work at the resonant frequency of the transducer, thus causing the ultrasonic transducer to stop or be damaged. In order to solve this problem, the equivalent circuit method is used to analyze the resonant frequency characteristics of the ultrasonic transducer, and a frequency tracking method based on FPGA is designed to track the resonant frequency of the ultrasonic transducer in real time. First of all, DDS is used to generate drive signals with adjustable frequency. Then a phase detection module is designed to detect the current and voltage phases of the transducer in real time. At the same time, in order to improve the tracking efficiency, frequency tracking is realized by PID algorithm. The experimental results show that the frequency tracking method designed in the experiment can quickly achieve frequency tracking, and the tracking error is kept within ± 2Hz and avoiding the problems of anti-resonance frequency tracking and locking loss during the tracking process. It has the characteristics of fast response speed and strong robustness.  
      关键词:ultrasonic transducer;DDS;resonant frequency;phase detection;PID;FPGA   
      9
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544550 false
      发布时间:2023-11-27
    • ZHANG Jianbing,CHEN Conglei,LIAN Yuanfeng,YAN Zexiao
      Vol. 22, Issue 11, Pages: 141-147(2023) DOI: 10.11907/rjdk.231704
      摘要:In order to meet the requirements of marine environment operation simulation and subsea installation applications, Vortex and Vega Prime virtual reality platform are combined to build a real-time dynamics virtual simulation environment. The virtual simulation training system for subsea manifold installation is designed and implemented by physical property modeling and geometric property modeling of working environment, manifold, rope, hook, guide pile, ROV. The experimental operation results show that the system has a highly physical realistic real-time dynamic simulation effect, achieving the entire process simulation of subsea manifold installation, with strong immersion and good interaction effects, and can meet the training needs of subsea manifold installation.Besides, simulation of subsea installation training for multiple joint control manifolds has been achieved, enhancing the realism of training and providing a highly reliable simulation platform for subsea operations.  
      关键词:subsea manifold;virtual simulation;Vortex;Vega Prime;training evaluation   
      13
      |
      2
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544707 false
      发布时间:2023-11-27
    • ZHAO Haonan,CHEN Xiaohong,SUN Fujia,LIN Xiaojia,HUANG Yongguo
      Vol. 22, Issue 11, Pages: 148-154(2023) DOI: 10.11907/rjdk.222370
      摘要:In the working process of ultrasonic scalpel, there are high requirements for the real-time and stability of the energy generator control system. Aiming at these requirements, a control system of ultrasonic scalpel energy generator based on DSP is designed. The system takes TMS320F28377D with main frequency of 200 MHz as the main control chip, and controls the power supply circuit, drive circuit, feedback circuit and man-machine interaction system through ePWM module, IIC module, SPI module and SCI module of the chip, and generate sinusoidal AC signals that drive the ultrasonic scalpel to work. Then, the control law accelerator CLA is called to carry out non-periodic sampling algorithm and PID control algorithm to complete the closed-loop control of the whole energy generator system. Compared with the MCU control system whose data operation and control functions are all completed by CPU, the control system using CPU and CLA together has faster operation speed and shorter control cycle. The power control strategy of constant output current is adopted to maximize the stability of the ultrasonic scalpel, and the parameters such as current signal and voltage signal of the system are displayed on the screen in real time.  
      关键词:ultrasonic scalpel;energy generator;DSP;embedded system;control system   
      11
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544396 false
      发布时间:2023-11-27
    • HU Hanwen,XUE Meng,GUO Hanming
      Vol. 22, Issue 11, Pages: 155-160(2023) DOI: 10.11907/rjdk.222351
      摘要:The main control system of the four-channel Raman spectrometer was designed according to the requirement of the acquisition of the material Raman spectrum irradiated by the laser at 532 nm, 633 nm, 785 nm and 1 064 nm. In this system, S11511-1106 planar array CCD and G14237-512WA linear array indium gallium arsenic image sensor of Hamamatsu company are used as photoelectric converters. STM32F407 is used as the control chip to realize the control of four-channel image sensor, four-channel laser and four-channel motor. FPGA is used to drive the image sensor, A/D analog-to-digital conversion and spectral data acquisition. The data is transmitted to the upper computer through the serial port for display. The experiment shows that the spectral signal of the upper computer is accurate, and the single or continuous acquisition of the spectrum can be realized. The voltage error affecting spectral imaging is stable within 1.5%, and the ripple coefficient is controlled within 2%. The system realizes the acquisition and transmission of four-channel spectral data, which can be extended to the design of other image components and has certain application value.  
      关键词:STM32;FPGA;CCD;indium gallium arsenic image sensor;multi-channel spectrometer   
      65
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544765 false
      发布时间:2023-11-27
    • JIANG Lihao,LIU Xingqiao,LI Changfeng,CHEN Hui,MA Teng,ZHAO Dean
      Vol. 22, Issue 11, Pages: 161-167(2023) DOI: 10.11907/rjdk.222161
      摘要:A modified Euclidean clustering segmentation method based on adjacent workpiece region recognition is proposed to address the issues of undersegmentation and oversegmentation in single workpiece segmentation in multi workpiece detection. Firstly, filter the collected point cloud directly to obtain the region of interest; Then, a region recognition algorithm based on outlier detection, normal vector angle, and density clustering is used to identify adjacent workpiece regions; Finally, combined with K-nearest neighbor based adaptive Euclidean clustering for point cloud segmentation. The results show that the proposed method has a segmentation accuracy of over 95%, can accurately segment the region of interest of a single workpiece from the scene, and the segmentation effect is relatively stable, which can lay the foundation for subsequent workpiece defect detection work.  
      关键词:machine vision;Euclidean clustering;point cloud segmentation;cluster segmentation   
      14
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544253 false
      发布时间:2023-11-27
    • XU Hongkui,GUO Wentao,LI Zhenye,GUO Xubin,ZHAO Jingzheng
      Vol. 22, Issue 11, Pages: 168-173(2023) DOI: 10.11907/rjdk.222339
      摘要:To solve the problem that the lightweight network structure is insufficient to extract effective semantic information from the feature map, and the design of semantic information and spatial detail information fusion module is unreasonable, which leads to the reduction of segmentation accuracy, designs a deep fusion hole pyramid and a semantic fusion module to further extract semantic information and perform deep fusion. In the backbone network, a dual branch deep fusion network is used. Through the designed deep fusion hole pyramid structure, further information extraction is carried out. In the process of spatial information and semantic information fusion, semantic fusion modules are used to gradually integrate deep level semantic information and shallow level spatial information in a bottom-up manner, thereby improving the performance of network semantic segmentation Finally, the designed network structure is tested on the urban landscape datasets, and the average intersection and merger ratio is 77.6%. Compared with other network structures, the segmentation results are effectively improved.  
      关键词:lightweight network;dual branch network;empty pyramid;information extraction;deep fusion   
      23
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544973 false
      发布时间:2023-11-27
    • CHENG Zehua,HAN Junying
      Vol. 22, Issue 11, Pages: 174-179(2023) DOI: 10.11907/rjdk.222427
      摘要:In the current mask recognition and detection task of YOLOv5 algorithm, when the light is dim, individual targets can not be correctly identified, and there are also problems such as large network parameters and slow running on mobile devices. Through adaptive feature fusion (AFF) of space separable convolution and depth separable convolution, a SDC (spatial deep convolution) module integrating attention is designed, which reduces the parameters of the network model, enhances the network feature extraction ability, and improves the detection accuracy of the model. The data enhancement methods of random occlusion and adaptive contrast are used to simulate the phenomena of occlusion or dim light in real scenes. SIoU Loss is selected as the frame regression loss function to improve the training speed and reasoning accuracy of the network. The experimental results show that compared with the original YOLOv5 algorithm, the mean precision (mAP) is improved by 1.3 percentage points, and the detection frame rate (FPS) is improved by about 140%, which better meets the accuracy and real-time performance of the mask detection task.  
      关键词:mask detection;detect frame rate;SIoU Loss;YOLOv5;dim light   
      32
      |
      6
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544971 false
      发布时间:2023-11-27
    • JING Liang,BAO Zhiyuan
      Vol. 22, Issue 11, Pages: 180-185(2023) DOI: 10.11907/rjdk.222506
      摘要:Aiming at the problem that the detection of obstacles by unmanned agricultural machinery in complex orchard environment is not real-time and accurate, an improved Ghost-CA-YOLOv4 orchard obstacle detection algorithm based on YOLOv4 is proposed. This algorithm selects the lightweight network GhostNet as the backbone feature extraction network, and introduces CA mechanism in GhostNet to fuse spatial information and channel information, so as to improve the target attention. Soft-CIoU-NMS algorithm is used to enhance the detection effect when the targets are dense or overlapping. To verify the effectiveness of the algorithm, the mAP value on the self-made orchard obstacle data set reached 90.75% after training in the Python framework. With the detection accuracy unchanged, the model size is compressed to 18.24% of the original YOLOv4, and the FPS is 45, which can better provide security for the operation of unmanned agricultural machinery.  
      关键词:deep learning;target detection;GhostNet;attention mechanism;Soft-CIoU-NMS   
      13
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544250 false
      发布时间:2023-11-27
    • ZHANG Tao,LI Chunxin,HU Ping,WANG Jiahui
      Vol. 22, Issue 11, Pages: 186-193(2023) DOI: 10.11907/rjdk.231738
      摘要:Online learning space provides a practical field for the smooth implementation of online learning. Learning behaviors reflect the path and process of learners' learning in online learning spaces, and group learning behaviours, as a specific manifestation of online collaborative learning process, have become a key area of concern for learning analytics to explore the path of learning behaviors of learning groups and the relationship between the path of behaviors and learning effects. Therefore, the article takes learning behaviors in the online learning space as the research object, and makes comprehensive use of content analysis, social network analysis, lag sequence analysis and correlation analysis to explore the overall social network structure of the group, the transformation mode and characteristics of learning behaviors, as well as the impact on learning effects. The results show that the social network structure of online learning groups is loose, and it is necessary to stimulate the intrinsic motivation of different learning groups; the group interaction behaviour is dominated by the shallow interaction structure, and it should be focused on the improvement of the interaction quality; the sequence of behaviours of different learning groups varies significantly and shows a certain regularity; and the group learning behaviors have a positive and significant impact on learning performance.  
      关键词:online learning space;different learning groups;learning behavior;behavioral sequence analysis;learning achievement   
      8
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545155 false
      发布时间:2023-11-27
    • TAO Kai,WANG Qiang,ZHENG Wei
      Vol. 22, Issue 11, Pages: 194-198(2023) DOI: 10.11907/rjdk.231011
      摘要:In traditional teaching method of virtual instrument technology course, there are problems such as lack of applicability and low student interest. For meeting the requirements of information technology majors under the background of new engineering, an online and offline mixed flipped classroom teaching method based on "three elements, three modules and two key points" was proposed. This teaching method could take the advantage of online and offline teaching resources. Teaching comparison experiments showed that the enthusiasm and sense of participation could be improved by this method, so as to a high learning efficiency.  
      关键词:virtual instrument;flipped classroom;online and offline;mixed teaching   
      10
      |
      1
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544321 false
      发布时间:2023-11-27
    • QI Jun,LIU Zhiqiang,LI Xinli,LIU Ping,REN Yangfu,ZHU Xiaolan
      Vol. 22, Issue 11, Pages: 199-204(2023) DOI: 10.11907/rjdk.231865
      摘要:In response to the large gap in students' proficiency and insufficient focus on teaching key and difficult points in the source areas, a scientific and reasonable course teaching design is carried out from three perspectives: the division of teaching key and difficult points, hierarchical teaching, and student warning, based on the process data generated by the online offline hybrid teaching of database application foundation courses. The questionnaire survey results show that students' teaching feedback is good, laying the foundation for more in-depth use of teaching process data for curriculum reform in the future.  
      关键词:mixed teaching;teaching process data;data analysis;instructional design   
      8
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545377 false
      发布时间:2023-11-27
    • WU Na,GE Min
      Vol. 22, Issue 11, Pages: 205-211(2023) DOI: 10.11907/rjdk.231688
      摘要:In the modern teaching situation, individualized teaching is limited by the description, evaluation, storage and use of students' personalized data, which becomes the main obstacle to the complete and large-scale implement of the educational concept of individualized teaching, and to a certain extent affects the education quality. The application of digital twin technology in the field of education can build a corresponding four-dimensional virtual model for the education process, and form a simulation analysis of physical entity status and behavior report, so as to provide teachers with a relatively perfect teaching simulation and feedback mechanism for the education process, so that teachers can carry out targeted teaching according to the individualized learning mode built for each student.  
      关键词:individualized teaching;digital twin;individualized learning model   
      6
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545051 false
      发布时间:2023-11-27
    • FENG Ling,FENG Bin,DUAN Xiqiang,ZHANG Lei,ZHANG Guofeng
      Vol. 22, Issue 11, Pages: 212-218(2023) DOI: 10.11907/rjdk.231807
      摘要:In response to the problems in the teaching of network security technology courses for computer majors in local undergraduate colleges and universities, based on the analysis of the demand for application-oriented talent cultivation, the current research status of this course teaching is analyzed, and its teaching philosophy is explored in the context of engineering education and new engineering. At the same time, the integration of schools and enterprises is developed to jointly explore teaching pain points, construct teaching systems, revise teaching outlines, reshape teaching content, mix teaching models, and joint teaching methods A series of reform measures such as comprehensive teaching evaluation. Teaching practice has shown that students' knowledge level and practical skills have been improved, providing reference and demonstration experience for teaching reform of other similar courses.  
      关键词:network security technology;curriculum reform;school-enterprise integration   
      5
      |
      4
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545121 false
      发布时间:2023-11-27
    • WANG Chen,LIU Yonggui
      Vol. 22, Issue 11, Pages: 219-225(2023) DOI: 10.11907/rjdk.231904
      摘要:Generative artificial intelligence, represented by ChatGPT, has sparked a new technological wave. Starting from the application of generative artificial intelligence, this paper explores a teacher's autonomous learning mode based on generative artificial intelligence. By analyzing the advantages of generative artificial intelligence in the field of adult autonomous learning, the feasibility of its application in teacher professional development is demonstrated. Focusing on the process of teacher self-directed learning, based on adult learning theory and combined with Zimmerman's self-directed learning theory model, this paper analyzes the assistance that generative artificial intelligence can provide in teacher self-directed learning from the perspectives of foresight, performance, and self-reflection. Generative artificial intelligence can help teachers set learning goals, provide personalized resources, and recommend learning strategies before learning, provide guidance, answer questions, and provide suggestions during the learning process, and assist teachers in evaluating learning effectiveness, conducting self reflection, and planning further learning actions after learning. Finally, the importance of generative artificial intelligence as an auxiliary tool was emphasized, emphasizing that teachers should actively participate in the autonomous learning process and recognize the possibilities and limitations of artificial intelligence.  
      关键词:generative artificial intelligence;ChatGPT;teachers' autonomous learning;autonomous learning mode   
      11
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545072 false
      发布时间:2023-11-27
    • LI Xiangju,CAO Fengping,XIE Xiujuan,WANG Mengxiao
      Vol. 22, Issue 11, Pages: 226-230(2023) DOI: 10.11907/rjdk.231652
      摘要:According to the construction standards of the "golden course", reform the Java programming course and construct a teaching resource that integrates "value shaping, knowledge transmission, and ability cultivation" to achieve the goal of cultivating the application system ability of applied undergraduate computer majors. In the process of teaching reform, selecting appropriate entry points, integrating ideological and political education into the curriculum, and providing value guidance to students; Emphasize the training of computational thinking and lay a solid foundation of professional knowledge; By designing practical cases that run through the entire course, students can enhance their understanding of computer application systems, equip them with system design and implementation abilities, and enhance their analytical and problem-solving abilities for complex engineering problems.  
      关键词:Java programming;Golden Course;Curriculum ideological and political;application system ability;computational thinking   
      10
      |
      2
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44545360 false
      发布时间:2023-11-27
    • XU Fengyu,WANG Qiang,YIN Haitao,DING Jie
      Vol. 22, Issue 11, Pages: 231-234(2023) DOI: 10.11907/rjdk.222404
      摘要:The in-depth development of new technologies such as artificial intelligence and cloud computing in the context of intelligent manufacturing has promoted the rapid transformation and upgrading of China's industries, and has also put forward new requirements for talent cultivation in information universities. The integration of industry and education is an important way for the development of information universities in the context of new engineering, and an important means of effectively promoting the integration of education chain, talent chain, industry chain, and innovation chain. Based on the demand for the construction of the artificial intelligence industry education integration major and the new concept of "innovative, comprehensive, and full cycle" engineering education, this paper proposes the overall idea and teaching reform plan for the construction of the artificial intelligence major in information universities. The aim is to strengthen the construction of the artificial intelligence industry education integration major and cultivate high-quality professional talents engaged in the design, development, application, and management of artificial intelligence systems.  
      关键词:artificial intelligence;integration of production and education;professional construction;curriculum system   
      12
      |
      3
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544664 false
      发布时间:2023-11-27
    • LYU Jia,ZENG Mengyao,PENG Gangjian
      Vol. 22, Issue 11, Pages: 235-240(2023) DOI: 10.11907/rjdk.222284
      摘要:This study aims to explore the pathway for curriculum development in the field of artificial intelligence (AI) at local universities, taking the intelligent medical engineering program at a medical university as an example. By starting from other distinctive disciplines at local universities, building upon the foundation of computer science courses, enhancing interdisciplinary backgrounds, and focusing on professional competence and practical skills, the study aims to establish a coherent and effective curriculum system that integrates AI with other disciplines, in order to meet the development needs of both the country and the local region. The research adopts methods such as literature analysis, on-site investigations, and expert interviews to analyze and refine the curriculum system of the intelligent medical engineering program. The results indicate that through well-designed and integrated courses, students' professional competence and practical skills in AI can be enhanced. Therefore, this study provides valuable references and guidance for local universities in cultivating talents in the field of AI.  
      关键词:artificial intelligence;local universities;curriculum system;cultivation of talents   
      6
      |
      2
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544942 false
      发布时间:2023-11-27
    • CHEN Mengting,XIONG Shixun
      Vol. 22, Issue 11, Pages: 241-245(2023) DOI: 10.11907/rjdk.222331
      摘要:Professional courses are the basic carrier of ideological and political education construction, which are of great significance to realize the goal of "three-integrity education". At present, the ideological and political elements of professional courses are not deeply explored, and the effect of ideological and political education is not obvious. The collaborative education mechanism of ideological and political and professional certification of the course “wireless positioning technology” is designed based on the professional certification standards of engineering education and combined with the discipline and professional characteristics. Carry out ideological and political teaching reform from the teaching content, teaching method and evaluation form, and sort out the ideological and political education elements in the curriculum comprehensively and reshape the teaching content and teaching process, so as to achieve the purpose of education imperceptibly.  
      关键词:ideological and political course;wireless localization technologies;professional certification;collaborative education;teaching reform   
      7
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544400 false
      发布时间:2023-11-27
    • YAN Dahu,YAO Jing
      Vol. 22, Issue 11, Pages: 246-252(2023) DOI: 10.11907/rjdk.222432
      摘要:Relying on the interdisciplinary advantages of STEM education to develop students' engineering design thinking has become the core goal pursued by both STEM education and engineering education. However, in current STEM teaching, the teaching mode based on engineering design mostly follows a rigid process, resulting in students staying at the level of mechanical operation and simple imitation, lacking innovation awareness. To address this practical dilemma, based on the analysis of the connotation of engineering design thinking and related research reviews, this paper clarifies the internal relationship between engineering design thinking and engineering design, explores the advantages of integrating reverse engineering and engineering design, and combines engineering design, STEM subject content, and reverse engineering teaching methods to propose a STEM teaching model for cultivating engineering design thinking. This model aims to help students establish a top-down perspective on engineering exploration, enabling them to perceive problems, solve challenging tasks, and achieve product innovation in the internal and external dual cycle of 'need to know' and 'need to do'. At the same time, a handheld small fan transformation case was designed and implemented based on this model, verifying its positive promoting effect on the cultivation of students' engineering design thinking.  
      关键词:engineering design thinking;reverse engineering;engineering design;STEM teaching model   
      6
      |
      0
      |
      0
      <HTML>
      <L-PDF><WORD><Meta-XML>
      <引用本文> <批量引用> 44544662 false
      发布时间:2023-11-27
    0