This PDF file contains the front matter associated with SPIE Proceedings Volume 12250, including the Title Page, Copyright Information, Table of Contents and Committee list.

Most of the crude oil produced in China is easily condensable with high viscosity, and the pipeline transportation method is mostly heating transportation. The energy consumption of this transportation method mainly comes from the heating furnace and the pump. To reduce energy consumption and find the optimal steady-state operation scheme of oil pipelines, this paper takes a crude oil pipeline as a practical case based on Matlab software, establishes a mathematical model for optimal operation of oil pipelines, and uses a genetic algorithm to solve the model. The results show that the genetic algorithm is featured by fast convergence rate and short optimization time, and can obtain the optimal scheme. Compared with the actual operation scheme, the optimal one can reduce energy consumption by 13.03%, which indicates that the model and method established in this paper have strong applicability.

Nowadays, many college students are lack of proper self-awareness, and unclear about their future, which reflects the conflicts between traditional college careers guidance and careers guidance needs of the present-day college students. With the popularization of mobile Internet technology, the campus career planning guidance should also gradually transform to informationization. Based on the needs of professional career planning guidance of college students, this paper discusses the design creativity of a professional career planning APP, sorts out the design ideas of this APP, hoping to help college students establish overall professional career planning system of four years in the college, and help them to strengthen their professional abilities and comprehensive capabilities. Through the research of this article, it is expected to provide new design ideas and reference direction for the design of professional career planning APP for college students.

With the continuous development of diagnosis, prediction and systems management technologies for complex pipeline systems and driven by the context of Industry 4.0, digital twin technology has become a research hotspot in the field of smart manufacturing and intelligent operation and maintenance of pipeline systems. Digital twin can be defined as an adaptive model of complex physical systems. The prospect of digital twin and its impact on society is brought closer to reality with recent advances in multiphysics simulation, artificial intelligence, big data cybernetics, data processing and management tools. With a case study on Sino-Myanmar pipelines, the author uses artificial intelligence models on the basis of digital twin technology to predict the pipeline transmission demand to make sure the smooth operation of pipelines with large topographic fluctuations and geological complexity. The objective is to achieve intelligent operation such as real-time autonomous optimization of pipeline operation and control of safety warning, thus realizing data analysis and intelligent diagnosis, as well as providing a reference for the application of demand prediction of complex pipeline networks and system health management in digital twin technology.

With the continuous development of Internet technology, the demand for network applications continues to increase, and people urgently need innovation at the network technology level. However, due to the highly coupled characteristics of traditional network architecture forwarding and control, it is difficult for current network equipment to provide highquality services for current network requirements. This paper utilizes the characteristics of transfer control separation, centralized control and programmability under the SDN network architecture. The application advantages of SDN technology in 5G communication network are analyzed, and a three-layer switch based on OpenDaylight controller that can support 5G communication technology is designed and implemented. This has important implications for eliminating the differences in the underlying physical devices of the network.

Load forecasting is an essential part of the energy field. The excellent performance of the prediction algorithm can directly reduce the operation and maintenance cost of the equipment. Therefore, making load forecasting more accurate is a concern to energy experts. In the types of load forecasting, short-term load forecasting (STLF) is usually used to test the accuracy of new algorithms. In the previous papers, amounts authors provided various methods to improve load forecasting accuracy. In recent years, STLF based on the neural network has been proposed by researchers. But few papers mentioned combined energy disaggregation technique and neural networks to improve the forecasting model. Therefore, this paper proposed a new method that combined energy disaggregation and long-short term memory (LSTM) to enhance the accuracy of traditional STLF. In this article, the author will introduce a non-intrusive load monitoring tool kit (NILMTK) for getting each household appliance consumption from the total consumption and then predicting the future load value of each appliance based on its load data. Finally, sum the predicted values to get the result. Through experiments, the results of our proposed new method reduce the root mean squared error (RMSE) of traditional forecasting models by 7% and the mean absolute error (MAE) of it by 15%, respectively.

With the rapid popularization of online education, the traditional mode of volunteer teaching has gradually become unable to meet the actual needs of volunteer teaching activities. With the objective of “promoting poverty alleviation through education”, based on needs analysis and combined with collaborative theory, the U-V-S “Internet + Volunteer Teaching” model has be established, and the "Cloud Volunteer Teaching" APP has been developed with the main functions of user management, resource management and cloud volunteer teaching. After testing and application, the APP runs stably and functions properly, which can effectively make up for the shortcomings of the traditional mode of volunteer teaching, and play a significant role in enriching volunteer teaching modes and boost educational poverty alleviation.

Spiking Neural Network (SNN) is a newly high bionic neural network with great potential in the future. However, the relevant research is not enough and there is no reference to the research direction at present. Therefore, we'd like to provide valuable references to researchers and enhance the understanding of SNN through this study. We analyzed 4005 papers and 1318 preprints about SNN from databases including Web of Science, Advanced Intelligent Systems, arXiv, and ResearchGate. By using visualization tools named CiteSpace and wordcloud, we made multifaceted exploration and analysis including keywords, institutions and countries, authors, journals, references, and preprints. We have explored the possible future research hotspots and trends of SNN, and we also explored the most influential authors, institutions, and journals in this field.

In order to adapt to the new changes in the network era, to solve the problems of the high cost of navigation radar equipment, the high cost of navigation teaching and training and maintenance, and to adapt to the grid and decentralized learning habits of Modern Society, a mobile device system based on a maritime radar simulator is required. FURUNO radar is widely used in ships due to its stable quality and low price. This paper takes the mainstream FURUNO FAR- 2827 radar as the research object, designs and develops a set of navigation radar simulation training system based on Android system. Based on the Douglas-Peucker algorithm, the coastline data is simplified, and the circular scanning method is used to simulate the radar image, and the coastline echo is generated by using the scanning line and the coastline intersection algorithm.

In order to reduce the network load and the access delay of web users, caching and prefetching technology are considered to be the main methods to solve the problems of web access delay. At present, the research on caching and prefetching technology often only emphasizes the improvement of replacement strategy or prefetching algorithm, lacks the research on prefetching method, and the improvement of access delay is not ideal. Combined with caching and prefetching methods, a hybrid Markov caching and prefetching model considering the change of user browsing trend over time is established. Aiming at the shortcomings of the current web page prefetching methods, the model uses the hierarchical clustering method to find out the time-varying interval of user browsing behavior. Each time interval uses a Markov chain to predict and maintain the freshness of information. The experimental results show that the proposed method can improve the cache hit rate and reduce the network access delay.

The existing antiaircraft fire system is a multi-level system, including high-precision long-range missiles and low-precision short-range antiaircraft artilleries. When firepower nodes are destroyed by the enemy or locked by missiles, it is necessary to position the target with other available collimating equipment. Therefore, it is of great significance to study how to achieve high-precision cooperative collimating and locating effect with direction finding and ranging data of lower precision. The problem can be solved by the principle of extended Kalman filter, and improved by an introduced feedback link in the simulation process. It has proved that this method is conducive to gradually narrowing the data error to stability in different time spans, which is applicable to this coordinative problem and promotes the effectiveness of air defense combat.

Software defined networks (SDN) is a novel network architecture, which decouples the control layer and data layer of the original network architecture. This network architecture utilizes controller to control data transmission. With the expansion of network scale, a single controller is unable to manage the whole network, while multiple controllers need to be deployed. Consequently, how to deploy multiple controllers appropriately is a hot topic in SDN. Furthermore, many intelligent algorithms have been applied to solving this problem. In this paper, we employ Cuckoo Search Algorithm (CSA) to improve the ability of multi-controller deployment. In order to improve the local search ability of the algorithm, the standard CS algorithm is improved in this paper. Firstly, weight coefficient is introduced to adjust the convergence speed of the algorithm. Secondly, all individuals sort according to their function values, after sorting the population can be divided into two parts. The experimental results demonstrate that the modified CSA performs better than other algorithms in multi-controller deployment.

Turbo code is a channel coding and decoding technology approaching Shannon limit. As a low complexity Max-Log-MAP (MLMAP) decoding algorithm, it is widely used in industry. At low SNR, the decoding performance of MLMAP algorithm of Turbo code is relatively poor, so it needs to be improved. The work of this paper improves the original decoding optimization algorithm of MLMAP algorithm, and proposes MLMAP-DSF algorithm. The optimal parameters of the algorithm are obtained through a large number of simulation experiments. The decoding performance simulation shows that MLMAP-DSF algorithm can improve the decoding performance of the original algorithm by about 0.6dB.

Embedded system includes embedded operating system, application software and support software, which is widely used and has high technical content. Take the single-chip computer as an example, the memory involves the control programs, but it does not involve the micro-operating system. The basic functions are no different from PC, but the advanced functions are obviously different from PC. The PC mainly stores the software on disk, but the embedded systems embed the software in FLASH memory. This paper mainly formats the data structure in the computer system, constructs the organizational framework of the embedded system, develops the embedded software, and deeply studies the application of the embedded software in the computer system.

The traditional A* algorithm in the field of meteorological navigation performance and accuracy is insufficient, and cannot meet the use of ships in navigation habits. This paper provides a feasible ocean weather routing optimization algorithm, in order to improve the computing performance and satisfy the navigation of ships used as optimization object, with ship characteristics and voyage forecasting meteorological as constraint conditions, sea conditions to calculate the accurate time saving fuel recommended route as the ultimate goal, through dynamic adding neighbor relationships between each grid, The ship dynamic model based on meteorological sea state influence factor was constructed, the whole route calculation process in A* algorithm was optimized, and the effective solution to the problem was realized by smoothing Bessel function. The experiment and the captain's actual use prove that the meteorological navigation is optimized by optimizing the calculation performance and smoothing process of the route.

Aiming at the shortcomings of poor platform universality, long preparation time and high safety fire risk in the experiment of individual rocket weapon, a kind of general launching platform based on magnetic induction sensor is designed. The height adjustment mechanism of launching platform adopts worm gear and worm drive mechanism. The direction adjusting mechanism adopts tooth arc and gear connection mode. The rotating part adopts the worm gear and worm speed reduction mechanism. By using the method mentioned above, the problem of freely adjusting the launching platform in three degrees of freedom is solved. The universal launching target of individual rocket weapons is realized. The electromagnetic firing is realized by using magnetic induction sensor and electromagnet, and the firing safety is improved. The design of the launch platform has been verified in the actual experiment.

Nowadays, ecotourism is becoming more and more popular, and it is more and more important to monitor and restore ecological resources in nature reserves. Applying GIS technology to the monitoring of ecological protected areas and providing managers with more detailed information and data can enable better protection of the ecological environment and more scientific and standardized management. The GIS system can centralize, store, operate and display geographic reference information, and can provide technical support for solving resource and environmental problems and ensuring sustainable development. In this paper, an evaluation index system of ecotourism suitability is constructed according to several indicators that affect the quality of the ecological environment, and a hierarchical structure model is used to accurately evaluate the ecological environment.

SRGAN applies generative adversarial networks to super-resolution reconstruction, and the advent of this algorithm greatly facilitates image reconstruction based on deep learning. It introduces batch normalization for stable training, and although batch normalization can facilitate training, it destroys the original contrast information of the image and affects the quality of image reconstruction. For this problem, this paper proposes an super-resolution algorithm based on improved SRGAN, which removes the batch normalization in the residual block of the generative network to avoid its bad influence on the reconstructed image and improve the quality of the reconstructed image for better. In addition, the total variation is introduced in the loss function of the discriminative model to constrain the gradient change, which serves to stabilize the network training and accelerate the convergence. Experiments of the proposed super-resolution algorithm on a large number of public datasets show that this method achieves better results under various evaluation metrics compared with other similar algorithms.

With the rapid growth of emerging vehicle applications, the great demands on low delay and energy of the vehicles are driven, vehicular edge computing (VEC) has been proposed to enhance the computation capacities at the edge of the vehicle. However, traditional VEC severs use fixed edge nodes to offload tasks making the system cost high greatly. To alleviate this issue, we propose a method which utilizes the idle resources on intelligent vehicles to assist edge computing offloading in the heterogeneous network. In order to minimize the system cost (the total latency and energy), we introduce deep reinforcement learning (DRL) and propose an improved Q-learning algorithm to jointly solve the task offloading and processing problem, and make the optimal offloading decision. The simulation results show that the improved Q-learning algorithm can effectively reduce the total system cost and improve the quality of service of VEC.

In recent years, with the continuous deepening of research on artificial intelligence, data analysis technology, and the Internet of Things, digital twin technology has gradually become a basic subject for studies around the globe and has been widely used in construction, health care, and automotive industries. To solve the increasingly complex problems of regulation and optimization in the oil and gas pipeline industry and accelerate the construction of the digital twin of the pipeline, this paper, via introducing the development of digital twin technology, study the context awareness, autonomous decision making, and adaptability of digital twin technology used in the oil and gas pipelines, and analyzes the prospect of the application of digital twin technology in pipelines from six aspects, namely pipeline integrity management, dispatching optimization, asset management, life cycle management and so on. Finally, the energy consumption optimization is analyzed by taking the China-Russia eastern pipeline as an example, which provides reference for the application of digital twin technology in pipeline system.

Physical education is an important part of education and the carrier of education. Chi Jian, president of Beijing Sports University, put forward the basic concept of "returning sports to education", and the "Thirteenth Five-Year Plan for Youth Sports" clearly pointed out: "Adolescents' health is an important part of a healthy China. It is the hope and source of Chinese competitive sports, and the foundation and guarantee for the realization of the Chinese dream. "In recent years, with the gradual enhancement of people's health awareness and the rapid development of the sports industry, sports majors in colleges and universities have also continued to emerge, and the number of enrollment has continued to increase, which has led to the cultivation of sports talents in high schools, and how to develop scientifically and sustainably. The cultivation of high school sports talents is a problem that the school has been researching and exploring.

In the official Android documentation, although it has described various configuration elements in the configuration file, some descriptions do not give accurate semantic definitions, and some descriptions do not show accurate relationships between related configuration elements, namely At present, there are uncertainties and ambiguities in the description of natural language in the document, which will cause some differences in the understanding of developers, and may cause developers to improperly configure the App to produce some security vulnerabilities. Therefore, this article mainly adopts the fuzzing-based security configuration rules research framework (ARMG) based on reading official documents, that is, for the existence of some configuration files with fuzzy semantics, the Fuzzing test method of multi-dimensional strategy is used to construct test cases. Analyze the test results to further clarify the rule definitions of related configuration items, and then use MobSF to statically analyze the sample set, and observe the compliance of the samples and the effectiveness of the rules by parsing, outputting logs, and comparing them. The final rules are used by designers or developers to improve development efficiency and better avoid software security vulnerabilities.

With the advancement of technology, unmanned method has become a promising approach for data collection from the ubiquitous Internet of Things (IoT) in the power industry. However, making autonomous obstacle avoidance on unmanned aerial vehicles (UAVs) is a challenging problem. It is much harder when the UAV is small since only very lightweight sensors can be carried on. In this paper, we focus on this problem and propose a novel model through deep reinforcement learning to achieve autonomous obstacle avoidance for UAVs. Different from prior similar works, the proposed model adopts the policy iteration-based method and requires only raw RGB images obtained by a monocular camera. Simulation results show that the proposed model performs better when compared to the other models based on value iteration.

With the rapid development of the Internet and the increasingly complex and diverse network environment, it is of great significance to quickly and accurately obtain the topology environment of the target network for optimizing network management, locating network resources and ensuring network security. Most of the existing network topology detection methods are implemented based on the SNMP protocol, or use the Traceroute-like method for active detection. However, in a non-cooperative environment, that is, the router disables SNMP permissions or prohibits responding to probe packets to shield Traceroute probes, network probes will face problems such as missing information and blocked probes. Based on the original SNMP protocol detection, this paper adopts RIP, OSPF and other routing protocols to supplement the detection method, which greatly reduces the number of undetectable nodes. In addition, for a small number of router devices that are still undetectable, the Traceroute directional breakthrough method using the routing table as a clue is used to bypass these undetectable nodes, complete the missing information in the network topology, and further optimize and improve the detection results.

Chinese meteorological alert text issued by meteorological authorities needs to be free of spelling errors. Automatic spelling error correction can find text errors and give correction suggestion. Most existing studies focus on open domain text correction such as news etc., however, the methods for vertical domain text correction such as meteorological alert text has not been well studied. In this work, we utilize the template feature of meteorological alert text and propose error oversampling strategy to enhance the correction model training. As for the correction model, we use multi-task learning to train the correction model by accounting error detection and error correction simultaneously. Experimental results on real-world alert texts show that our proposed method is significantly better than the baseline, exceeding the baseline by 3% on F1 measure.

The errors of inertial measurement unite (IMU) mainly include gyroscope’s noise, the white Gaussian noise in gyroscope’s bias, accelerometer’s noise and the scale factor of external acceleration. In this paper, an efficient estimation method of these errors is proposed, which do not need turntable and benefits low computational complexity. First, mathematical models are given in order to estimate the values of gyroscope’s noise, the white Gaussian noise and accelerometer’s noise. Then, several values of the scale factor are obtained in real time by setting partial-differential equation of yaw’s error with respective to the scale factor to zero, and the value of scale factor is further set when orientation’s errors have the minimum value. Thus, the update of orientation could be realized only by optimizing the value of scale factor, and it is accurate as the affection of estimation errors of IMU’s errors on orientation’s accuracy could be offset each other. Experiment results of a body’s orientation with an IMU show pitch, roll and yaw deviate from reference values only by 0.006 °, 0.005 ° and 0.5 ° respectivity, which verify the feasibility of the efficient estimation method.

This article researches on a real-time positioning and tracking system based on spot detection in order to measure deflection of bridges with high frame rate and accuracy, and introduces a method of adaptive threshold for locating the centroid of a spot. The efficient threshold can be captured by the way and then helps to search the ROI as well as to segment the image, which makes the system can operate in most environment under various lighting conditions. The system has reached accuracy level of 𝟏𝒎𝒎 and frame rate of 𝟏𝟎𝟎𝒇𝒑𝒔(frames per second) in simulated experiments, which prove the system is available in measurement for deflection of bridges.

In view of the problems of the traditional network architecture based on the boundary, the zero-trust architecture has undoubtedly become one of the best solutions to the security challenges. But as far as zero-trust technology is concerned, there are problems such as the control center or database being attacked still exits. This paper discusses and restructure the internal structure of the basic framework of the zero-trust system through the application of endogenous security ideas. By introducing a heterogeneous redundancy mechanism, the zero-trust security model achieves endogenous security gains, and its system structure has universal applicability, it can intensively achieve the functions of network service, reliability guarantee and security defense.

SDN (Software Defined Network) is a network architecture that meets people's needs for network service quality, and can realize more flexible and efficient network traffic monitoring technology. However, the environment configuration on the SDN network is very troublesome, and the installation of multiple applications into the environment may conflict. In order to solve the above problems, Docker came into being. It is an open source application container engine that implements some functions similar to VM (VMware Workstation) and provides more computing resources to users. In this paper, Mininet is used to build SDN data center network topology, and Docker is used to realize virtualization. This paper uses the Ryu controller to obtain the flow information of the switch interface through the OpenFlow protocol, so as to obtain the load status and remaining bandwidth of the network link, which has important application significance. The research results of this paper can be directly applied to various applications such as SDN congestion control and load balancing.

With the rapid development of Internet of things industry, the Internet of things engineers and technicians were provided a very broad career development prospect. The construction of professional Internet of things system integration training room in higher vocational colleges would improve students' professional practice ability and broaden their employment channels. The construction ideas and objectives of the Internet of things system integration training room were expounded. According to the structure of Internet of things perception layer, network layer and application layer, the construction contents of the Internet of things system integration training room were introduced detailly, which were provided a reference for the construction of Internet of Things training room in higher vocational colleges.

The rapid development of information technology makes the network environment more complex and changeable, and also brings more complex network attack methods, which makes the power monitoring network face more challenges. Under such circumstances, deep learning techniques are increasingly becoming an important part of the attack detection field. Deep learning technology can make self-adaptive judgments on different network attack methods, realize the detection of network attacks, and then improve the accuracy of attack detection technology. This paper builds a power monitoring network attack detection system based on the convolutional neural network and deep neural network in deep learning, which can accurately identify network attacks and reduce the false alarm rate of the system, thereby ensuring the smooth operation of the power monitoring network.

The importance of the information transmission control technology in the computer electronic information system is self-evident. To play the function of the computer electronic information transmission control technology, we first need to use the electronic information transmission and communication to control the information transmission process between different devices. When the electronic information transmission link can be guaranteed, the operation of the computer electronic information system is also very stable. In addition, it is necessary to avoid the problem of poor transmission quality in order to unify the interfaces of different platforms. Therefore, this paper mainly studies the application of the information transmission control technology in the computer electronic information system by analyzing the information transmission control technology and designing the cross-platform methods of the information transmission.

Point cloud denoising is currently still an attractive topic in computer vision. The main challenge is how to preserve the neat geometric features of 3D models while suppressing surface noise effectively. In this paper, we propose a cascaded approach which sequentially includes normal filtering, feature extraction, multi-normal estimation, and points updating to address this concern. The proposed approach keeps robust and has outstanding feature preserving performance by using the multi-normal strategy and incorporating a new feature-recognized trilateral weight. Experiments on synthetic and real scanned point clouds demonstrate that our approach can preserve various geometric features while suppressing noise reliably and is more competitive than the compared methods.

COFDM technology has been applied in wireless transmission system of moving remote control target. The basic principle, advantages and disadvantages of coded orthogonal frequency division multiplexing (COFDM) are analyzed. A wireless transmission system of moving remote control target based on COFDM self-organizing network is designed to meet the requirements of low delay and high image definition. The coverage capability and transmission performance of the system are analyzed. The results show that the self-organizing network technology based on COFDM can completely meet the wireless transmission requirements of the existing moving remote control targets. The method has been validated in the sports car experiment, which lays a foundation for the following remote control target cluster communication application.

Software Defined Network (SDN) redefines and abstracts the functions of the control layer, forwarding layer, and application layer of network devices, making network device software programmable, better supporting dynamic control of network traffic, and making up for the natural defects of traditional networks. Now it has been adopted in many application fields. Therefore, the security of SDN has also received more and more attention. The research on self-secure SDN has become vital important. Self-secure SDN can ensure its own security, not only actively detect network attacks, but also effectively predict and defend attacks. This paper designs a self-secure system based on software-defined network. The system mainly includes three functional modules: 5G user classification, application identification and abnormal traffic detection. Through the system designed in this paper, we can not only view the traffic and user information of the network loaded with this software-defined network-based self-secure system, but also enable or disable these traffic and users. After testing, the system designed in this paper is of great help to the improvement of the security of a software-defined network.

Retrieval strategy is constructed based on incoPat, the index of patent citation frequency is used to mine the core patents in the field of global edge computing, and core patents are analyzed by backward and forward citation. Results show that USA is the main application country and source country of global edge computing technology; The core patent US20110075675A1 ranks first with an average annual citation frequency of 60.6 times. Its forward citation subjects include IBM, Cisco, Huawei, NEC, Samsung Electronics and other enterprises, and there is an innovative competition relationship; the evolution route of edge computing evolved from early edge call and session management to packet routing and wireless bearer identification, and finally to intelligent routing and wireless device monitoring strategy.

The multiphase flow transient model is used to calculate the change of the flow medium parameters along the pipeline aiming at the characteristics of high sulfur content and easy corrosion and perforation of the associated gas pipeline in oilfield A, and the Apriori association rule mining algorithm is used to determine the correlation between the flow medium parameters and the corrosion rate. Three corrosion degree judgment plates are established respectively for liquid velocity -liquid holdup, temperature-PCO2/PH2S and liquid velocity -PCO2/PH2S, and the plates are verified by on-site pipelines. The research result shows that the corrosion degree of the pipeline is consistent with the corrosion degree found by using the three kinds of plates. So the plates in this paper can be used to judge the corrosion degree of pipeline.

In the relationship research on Chinese basic education knowledge map, remote supervision is usually adopted to solve the problem of training corpus shortage, and use the neural network model to guarantee the construction of relationship model, so as to ensure the accuracy of relationship research. At the same time, in the practical research, the effect of alleviating the introduction of wrong labels by the remote supervision will be considered, so in the model design of the relationship research, the semantic information of the bidirectional context can be collected with the help of the two-way gating cycle unit. This research will study the relationship between Chinese basic education knowledge map on the basis of Chinese information technology textbooks and Baidu Encyclopedia, and the purpose is to have a deep understanding of the concern, so as to provide a guarantee for the establishment of Chinese basic education knowledge map, so that the relevant work can be carried out smoothly.

With the rapid development of global shipping industry, intelligent recognition of container number has become a hot spot of application. However, the containers are usually outdoors, which often can't reach the normal recognition environment of characters. For complex scenes such as corrosion, tilt, blur and uneven characters, the traditional text recognition methods are not ideal due to high time-consuming and multi-stage reasons. In this paper, a new method for recognition of container number of port is proposed. Based on the deep learning algorithm, the multi-layer neural network is used to identify features of multiple dimensions, enhancing the algorithm's ability to identify container number in complex scene. The method consists of four parts, the original container image is processed by the improved EAST network, AddMargin, CRNN network and calculation operation of check digit, the accurate container number can be obtained as the result. Through experiments, 3961 samples of container images were tested, and the accurate rate of recognition of container number could reach 94.14% (3729/3961), and the detection speed was 1.20 seconds per image.

In this paper, we address the characteristics of various associations implied between fake reviewers and fake reviews, and the ability to construct review network graphs with dependency relationships. Firstly, we use graph neural networks to extract association features between reviews, propose a method to improve graph neural networks based on TrustRank and propose a multi-feature detection framework. Then the graph neural network is used to aggregate neighborhood feature output embeddings and connect them with the paragraph vectors extracted by Doc2vec. Finally, the experiments show that the algorithm in this paper can detect more fake reviews and has a better check-all effect.

As the number of college students increases, the identity of college students gradually depreciates. The ensuing topic of "certification fever" also enables college students to enhance their value by obtaining various skill certificates. In recent years, the development environment for using WeChat mini-programs to develop various program requirements has become more and more mature. Tencent mini-programs have simple operations, take up almost no memory, and are born with the operation of WeChat. The scheme and design of the college student certification guide provides the basis for the subsequent developers. The applet is developed using a three-tier architecture, using HTML+CSS+JavaScript development language, and MySQL as the database. The WeChat applet transfers data to the backend interface in the form of json messages. The test guide applet for college students realizes the functions of various certificates classification, test information query, test forum and question bank, and realizes a friendly test guide user interface.

Shelf labels have always been an essential role in supermarkets and warehouses, but with the expansion of business scale, the complex management problems of ordinary price labels have also become prominent. In recent years, electronic shelf labels have been widely used, but most of them use 433M wireless communication or WiFi communication. This paper presents a shelf electronic label developed using Bluetooth protocol, which uses the ink screen module based on electrophoretic display technology as the label display. The complete electronic shelf label system includes three parts: server, Bluetooth gateway and electronic shelf label. Server communicates with Bluetooth gateway using MQTT protocol. This paper introduces the design and implementation of electronic label and Bluetooth gateway.