Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259301 (2023) https://doi.org/10.1117/12.2673262
This PDF file contains the front matter associated with SPIE Proceedings Volume 12593 including the Title Page, Copyright information, Table of Contents, and Conference Committee Page.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259302 (2023) https://doi.org/10.1117/12.2671500
A coverage optimization method based on an improved sparrow search algorithm (LSSA) is proposed for the coverage problem arising from the initialization of wireless sensor networks. Firstly, the good point set method is used for population initialization to make the sparrow individuals uniformly distributed, and the algorithm can effectively avoid falling into the local optimization. Secondly, a nonlinear convergence factor is proposed to constrain the proportion of producers and scroungers, which ensures the diversity of the population during the search process and improves the solution accuracy. Finally, the location update method of producers is improved, and the algorithm’s convergence speed and optimization performance are improved by balancing global search and local search. The simulation results show that the improved sparrow search algorithm effectively achieves the optimal node deployment and improves coverage rate and convergence speed.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259303 (2023) https://doi.org/10.1117/12.2671469
Credit assessment models are an important basis for financial credit institutions to determine whether to lend or not, so an efficient and accurate credit assessment model is crucial for financial credit institutions. Traditional credit assessment algorithms do not take into account the noise problem caused by the massive amount of credit data, which greatly affects the time complexity and accuracy of credit assessment algorithms. In view of this, this paper proposes a credit assessment method based on information entropy and LSSVM. The method first uses information entropy to assign weights to feature attributes, then sets thresholds on them for feature extraction, and constructs an LSSVM model to evaluate credit data, so as to achieve accurate assessment of credit transaction risk. The experimental results show that the method can effectively reduce the time complexity of the algorithm and improve the accuracy of prediction
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259304 (2023) https://doi.org/10.1117/12.2671671
To solve the quantitative problem of enterprise innovation capability, a data driven quantitative method of enterprise innovation capability is proposed. Firstly, it analyzes and summarizes seven factors which affect the innovation ability of enterprises; Secondly, the enterprise is adaptively divided into different data clusters by deep clustering method; Thirdly, a Gaussian mixture model is constructed to quantify the innovation capability of the evaluated enterprise. The proposed method adopts data mining technology and can provide reference for enterprise development.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259305 (2023) https://doi.org/10.1117/12.2671657
In the multi-parameter sequence in the industrial electrolyzer, in order to solve the problem that the traditional method is difficult to predict the nonlinear features and obtain the hidden feature information in the sequence, this paper uses the VARMA model to fit the multi-parameter features and combines the Time2Vec vector to embed the time form as the neural network. Augmented data sources for automated feature engineering and generalization of deep learning techniques; multivariate parameters were dimensionally reduced and KS tests were used to capture correlations in order to explore relationships between electrolyzers. The experimental results show that the model is superior to other comparative models in terms of computational efficiency, accuracy, and network structure, which verifies the effectiveness of its prediction in the multi-parameter field.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259306 (2023) https://doi.org/10.1117/12.2671519
In view of the increasing data volume and the increasingly difficult data analysis in the power industry, an intelligent and efficient analysis and mining framework for power big data is designed to quickly obtain valuable information. Analyze the overall framework of the power big data center, mainly including the service layer, verification layer, data source layer, and feature analysis layer. In addition, through analyzing the process of data mining, it is found that the business needs to be strengthened And realize expansion. The framework design of power big data intelligent analysis and mining technology mainly includes power market demand, customer analysis, high-performance data analysis, service system, data security governance and other modules. Through the analysis of an example of intelligent power big data mining, the analysis results show that the intelligent power data mining has good analysis effect and high mining accuracy
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259307 (2023) https://doi.org/10.1117/12.2671640
At present, the parallel computing theory based on spatial big data has problems such as difficult algorithms, difficult operations, and complex formulas, based on this, this paper proposes a p-Dot parallel computing model based on the traditional parallel computing model of BSP (Bulk Synchronous Parallel), and then tests the model effect by setting experiments. The results reveal that: (1) All curves are open up and have a minimum value. (2) The dataset with a capacity of 0.25GB is the benchmark dataset. (3) The expansion rate e(w) of the input data capacity of the model under different test procedures has a linear relationship with the expansion rate e(n* ) of the corresponding optimal number of machines. (4) When 𝑛→∞ in the partition equation p(n), p(n) tends to a certain value.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259308 (2023) https://doi.org/10.1117/12.2672772
As the key basic supporting platform of electric power enterprises, the data center often has internal attribute revocation, which seriously affects the efficiency of information attribute encryption. This paper proposes a method of information attribute encryption for data centers based on a hash algorithm. Then, update that data platform information, ensuring that the encryption method has low overhead and high efficiency. Extract the attribute of the data platform information based on a hash algorithm and encrypt the attribute of the information. The simulation results show that the proposed method occupies only 24% of the task process. The encryption time is relatively short, which verifies that the method has low overhead and high efficiency in the process of information attribute encryption and has a certain contribution value to ensure the information security of the data center.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259309 (2023) https://doi.org/10.1117/12.2671444
Most of the subway stations are below the groundwater level, so it is particularly important to do a good job in waterproofing. For the leakage disease of subway station structure, detection is the method and identification is the purpose. There are many types of urban subway station structural leakage diseases, and it is not easy to identify the diseases in all directions. Based on the summary and analysis of the common identification methods of seepage water diseases in subway stations, the crack digital images obtained in a non-contact way are taken as the research objects. Starting from the crack characteristics and the principle of digital image processing algorithm, the traditional algorithm is improved and optimized to obtain a detection algorithm more suitable for the digital image of concrete structural cracks, which is applied to the identification of structural cracks in subway stations. Compared with other manual methods, this method is more accurate and can save a lot of time and cost.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930A (2023) https://doi.org/10.1117/12.2671487
A predictive coding algorithm for image lossless compression is introduced. In the prediction stage, the algorithm uses the local change rate of the pixel value to adjust the prediction model adaptively, and in the coding stage, the error feedback technology is used to further reduce the information entropy of the error image. The simulation test results on standard images show that the performance of the algorithm is significantly better than the standard lossless compression algorithm. The compression algorithm we proposed uses the local change rate of the image in the decorrelation phase to improve the prediction accuracy, and in the coding phase, the algorithm uses error feedback technology to further reduce the error.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930B (2023) https://doi.org/10.1117/12.2671344
In order to improve the efficiency of marine garbage classification, this study first enhances and processes the data set, then uses ResNet50 network model and modifies its lowest layer of network, and finally obtains the accuracy in different cycles through training and verification. The results show that the accuracy of the network model trained in this study is as high as 96% under the most stable cycle.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Meng Xiao, Lixin Liu, Shanshan Tian, Ang Li, Wenqian Zheng
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930C (2023) https://doi.org/10.1117/12.2672641
Based on the trend research of market supervision scheduling automation, this study analyzes the function of establishing market supervision scheduling automation operation and maintenance supervision system, expounds the application value of market supervision scheduling automation operation and maintenance supervision system, puts forward the key points of market supervision scheduling automation operation and maintenance supervision system design, provides the key points of market supervision application scheduling automation operation and maintenance supervision system application, It is hoped that through the evaluation of dispatching automation operation and maintenance supervision system, the design and application of dispatching automation operation and maintenance supervision system will be strengthened, and a dispatching automation operation and maintenance supervision system suitable for the actual development of market supervision will be established, which will be helpful for the popularization and application of dispatching automation operation and maintenance supervision system in market supervision.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930D (2023) https://doi.org/10.1117/12.2672143
The task of Coverage Path Planning (CPP) is to generate a route that satisfies the condition of reaching every possible area of a specific room. The room is divided into grids whose sizes are the same as the target moving in the space. Basically, CPP algorithms are classified into classical algorithms and heuristic-based algorithms. This paper focuses on one of the heuristic-based algorithms, A*, and applies four heuristic mapping functions to generate the coverage path. Subsequently, every grid is considered as the initial point where the path is generated, and the resulting steps and the cost of the total steps are compared. Besides searching for the routes available to cover the whole map, the ultimate target of the algorithm under discussion is to select the best initial point among all of the grids. This step ensures that the valid path generated by the algorithm is the shortest or the most commercial result.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930E (2023) https://doi.org/10.1117/12.2671587
Support vector machine is widely used in various fields because of its excellent generalization performance. However, the selection of its parameters directly affects the accuracy of the final results. An improved sparrow search algorithm (ISSA) is proposed to optimize the parameters of support vector machines. The ISSA algorithm improves the original algorithm from three aspects: replacing random method with optimal point set initialization population, changing the explorer position update formula, and adopting adaptive mutation mechanism. The UCI standard data set was selected to compare the SVM optimized by ISSA algorithm with the original SVM, the SVM optimized by the genetic algorithm, the particle swarm optimization algorithm and the basic sparrow search algorithm, respectively. The experimental results show that the classification accuracy of the SVM optimized by ISSA algorithm is significantly improved, and the generalization performance is further improved.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930F (2023) https://doi.org/10.1117/12.2671490
In order to effectively supervise the wearing of safety helmets by construction personnel, the YOLOv4-tiny target detection algorithm is used to detect the wearing of safety helmets. A lightweight model with higher accuracy and less computation is designed for YOLOv4-tiny, which is more suitable for real-time helmet wearing detection. Firstly, G-Resblock is designed to replace Resblock to reduce the computational complexity of the model and occupy less computing resources. However, YOLOv4-tiny is prone to error detection or missed detection in complex work scenarios. In order to solve this problem, an attention mechanism is added to YOLOv4-tiny, the serial channel of CBAM is improved as a parallel channel, and P-CBAM is added to YOLOv4-tiny to solve the problem of poor model detection effect. The improved YOLOv4-tiny can better complete the helmet detection task.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930G (2023) https://doi.org/10.1117/12.2672186
With the in-depth development of science and technology, computer based multi disciplines have become the focus of research. Cloud computing is a virtual data information mode, which not only reduces the computational complexity of the server, but also saves a lot of time. Therefore, cloud computing has become the most important research topic. In the modern management of enterprises, accounting informatization is an indispensable part of improving the operating efficiency of enterprises, which will constantly improve financial management, efficiency and supervision. Cloud computing is a mode of virtualizing data through software services, which is a related method of service related data computing without related hardware facilities. Through cloud computing, enterprises can store business information in remote service terminals, which will continuously improve the efficiency of enterprise processing. At present, cloud computing has become a new way of accounting informatization, which has replaced the traditional way of accounting informatization. As data virtualization becomes the norm, cloud computing has become the main line of system services, which will be more conducive to resource sharing of accounting informatization. First, this paper analyzes the related concepts of cloud computing. Then, this paper analyzes the influencing factors of accounting informatization, including internal and external environmental factors. Then, this paper analyzes the advantages of accounting informatization. Finally, some optimization measures are proposed
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930H (2023) https://doi.org/10.1117/12.2671376
This paper focuses on the intelligent recognition of images in the Tiangong remote sensing image dataset and its interpretability analysis. In this paper, we classified the aforementioned dataset, retrained the Resnet-18 model on the training set, and then verified the results on the validation set with an accuracy of 97.9%. Furthermore, this paper presented an interpretability analysis of deep learning for intelligent recognition of the Tiangong remote sensing image dataset.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930I (2023) https://doi.org/10.1117/12.2671570
Aiming at the problems of high modeling complexity and low rendering efficiency of existing visualization methods of real meteorological cloud data, a 3D visualization method of meteorological cloud data based on adaptive far-field grid structure of region of interest is proposed. Methods The region of interest was extracted to generate an adaptive far-field grid structure, which was applied to cloud particle modeling. The fine resolution of the region of interest was kept, and the number of particles in other regions was optimized. Finally, the rendering of 3D cloud images was completed. Simulation results based on WRF model meteorological cloud data show that the above grid structure can speed up rendering and rendering on the basis of ensuring the rendering quality, and can better display the morphology and structural characteristics of real clouds.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930J (2023) https://doi.org/10.1117/12.2671658
In complex environment, the performance of traditional face recognition algorithm decreases greatly. In order to further improve the recognition accuracy of current face recognition algorithms, this paper proposes two face recognition algorithms based on improved convolutional neural networks through the analysis of the defects of traditional algorithms. Finally, we will build a new face recognition model to verify the effectiveness of the two new methods. The first method is to extract and classify face features by fusing convolution layer and pooling layer, train neural network by stochastic gradient descent method, recognize face by Softmax classifier, and finally solve the over-fitting problem by "Dropout" method. The second method is to use the network link structure of bisymmetric LetNet and DCT-LBP joint processing method to process the input image. The two algorithms have some similarities, and both can improve the accuracy of face recognition.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930K (2023) https://doi.org/10.1117/12.2671338
The rapid development of artificial intelligence has prompted the convolutional neural network (CNN) to process huge amount of data, which has caused a great burden on convolution operations. Therefore, according to the characteristics of the systolic array architecture, the acceleration structure of CNN is constructed by fusing it with CNN. Besides, it is optimized in practical application, and its effectiveness is verified. The experimental results show that in the broadcast architecture, the time required by the CNN acceleration architecture is at least 0.005, while the maximum throughput is 16.83, which is far higher than the acceleration architecture under the systolic array architecture. In the case of small change in the maximum frequency, the error rate is the same as that of the systolic array, which is about 3.62%. In the comparison of various methods proposed on the systolic array, the accuracy rate of CNN acceleration architecture is 94.7%, and the utilization rate is 81.95%. The correctness and effectiveness of the algorithm are proved. To sum up, the improved CNN acceleration structure based on pulse array optimization reduces the response time and meets the requirements of terminal calculation force, which is of high significance in practical application
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930L (2023) https://doi.org/10.1117/12.2672190
In the network geographic information system, through voice interaction, the operation can be made simple, convenient and effective. To this end, this paper studies the GIS map component technology to support voice interaction. Build the overall design of GIS map components, which includes three layers: function layer, data layer and map UI layer. The functional layer is the main layer for realizing voice interaction. After audio enters the functional layer, voice recognition must be performed first. After understanding the semantics, the mapping feedback is completed, and voice interaction is realized and supported. Experiments show that the recognition speed of the content designed in this paper is relatively fast, and the highest recognition rate is 98.5%, which provides functional component support for the information processing of geospatial information.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930M (2023) https://doi.org/10.1117/12.2671455
When the Visual-Inertial Odometry (VIO) is started, its Inertial Measurement Unit (IMU) lacks acceleration incentive, which will result in poor orientation estimation accuracy during initialization, or even initialization failure. Therefore, a visual priori map-assisted monocular location algorithm based on 3D spatial straight lines is proposed. Firstly, the monocular image data of the surrounding environment were extracted through the Line Segment Detection algorithm (LSD), and high precision 2D line features were selected according to the length of the line and the number of surrounding point features. The 3D spatial lines of the surrounding environment were obtained using the line and surface intersection method. Construct a visual prior map with 3D spatial straight lines. Secondly, the constructed visual prior map is used as the online monocular VIO pose estimation for the global map. Based on the straight-line feature matching algorithm and the 3D space straight line depth information as additional constraints, the 2D straight-line feature in the monocular VIO's current field of vision is matched with the 3D space straight line in the visual prior map. The matching results were used as global constraints to optimize the monocular VIO pose. Tests on EUROC and TUM common data sets show that the 3D spatial straight line based visual prior map can effectively correct the pose during the monocular VIO initialization stage. Compared with the VINS-Mono localization algorithm, this algorithm can effectively improve the pose estimation accuracy during VIO initialization and reduce the overall trajectory positioning error.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930N (2023) https://doi.org/10.1117/12.2671479
The era of big data and cloud computing has driven the rapid expansion of the number and scale of data centers worldwide, and the ensuing huge power consumption has put pressure on resources and the environment. Accurate prediction of data center power consumption can provide an important basis for current power management techniques, while effectively improving the efficiency of intelligent operation and maintenance of modern data centers. To address this problem, a server power consumption prediction model based on a combination of principal component analysis (PCA) and DeepAR is proposed in the paper. The model uses the time series of server power consumption and performance index data from the Zhengzhou Inspur data center to predict future moment power consumption, performs principal component analysis on the performance index, and inputs the effective principal components and historical power consumption data into the DeepAR network for prediction. The model is experimentally validated on all three server datasets, and the results show that the model outperform the DeepAR network model as well as other comparison models in terms of prediction. When compared with the DeepAR network, the MAPE of this model is reduced by 0.23%, 0.12%, and 0.05% on the data1, data2, and data3 datasets, respectively.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Electronic System and Intelligent Network Recognition
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930O (2023) https://doi.org/10.1117/12.2671298
At presen, object recognition task is troubled by its huge kinds of objects. In this paper, the SIoU loss function and YOLOv5 deep learning convolutional neural network are innovatively used to improve the training efficiency and recognition accuracy. Unlike the traditional bounding box regression loss function (e.g. Giou, Diou[1] , CIoU) , which only focuses on the distance between the prediction box and the ground true box, the size of the overlap area, and one or more of the aspect ratios, and sets the impact factor on this basis, the SIoU loss function also introduces Angle cost to fit the best regression direction, which makes the direction of bounding box regression more reasonable and improves the regression efficiency[1].In this paper, the defects of traditional loss function and the calculation method of SIoU loss function are introduced, and the performance between SIoU and CIoU is compared.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930P (2023) https://doi.org/10.1117/12.2671830
Sculpture design and creation is a highly complex art form, requiring designers to have high artistic quality, innovation and practical ability. Traditional sculpture design can meet the needs of people for sculpture design and use at a certain stage, but it requires more links from the initial conception of the designer to the final presentation of the work. The efficiency of sculpture design is greatly limited, and the production process of the work consumes more materials. For this reason, this paper establishes a digital sculpture design method based on VR technology. By analyzing the application basis of VR technology in digital sculpture design, a sculpture design model construction method based on eye-movement data collection and feature extraction is established. Secondly, the feasibility of VR technology application is further verified by analyzing the application advantages of VR technology application compared with traditional sculpture design. In addition, this paper establishes a VR-based digital sculpture design and construction process, and analyses the beneficial effects of VR technology on sculpture design. The research in this paper verifies the effectiveness and feasibility of VR sculpture design, which contributes to the diversified development of sculpture design.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930Q (2023) https://doi.org/10.1117/12.2671649
The influencing factors of aluminum electrolysis production process are complex, and current efficiency is an important evaluation index. In order to study the influence of various parameters on the current efficiency in the aluminum electrolysis production process, a LightGBM-TPE current efficiency optimization model was established in this paper. First, the production data is preprocessed, and the industrial parameters are fitted using the LightGBM prediction model. Then, to further increase the model's prediction accuracy, the TPE optimization method is used to optimize the LightGBM hyperparameters. Finally, the optimization of current efficiency is realized through Optuna combined with TPE Bayesian optimization algorithm. The experimental results demonstrate that the model is capable of accurately identifying the realization conditions and process parameters of high current efficiency in the production process, as well as providing a parameter control foundation for the effective operation of the actual electrolytic aluminum production, ultimately achieving the goal of power consumption reduction.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930R (2023) https://doi.org/10.1117/12.2671835
In recent years, the outbreak of the COVID-19 epidemic has posed a serious threat to the life safety of people around the world, which has also led to the development of a series of online learning assessment technologies. Through the research and development of a variety of online learning platforms such as WeChat, Tencent Classroom and Netease Cloud Classroom, schools can carry out online learning assessment, which also promotes the rapid development of online learning technology. Through 2D and 3D recognition technology, the online learning platform can recognize face and pose changes. Based on 2D and 3D image processing technology, we can evaluate students' online learning, which will identify students' learning state and emotion. Through the granulation of teaching evaluation, online learning platform can accurately evaluate and analyze the teaching process, which can realize real-time teaching evaluation of students' learning status, including no one, many people, distraction and fatigue. Through relevant algorithms, the online learning platform can realize the assessment of students' head posture, which will give real-time warning of learning fatigue. Firstly, this paper analyzes the framework of online learning quality assessment. Then, this paper analyzes the face recognition and head pose recognition technology. Finally, some suggestions are put forward.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
YuJing Wang, RuiDa Ye, Tian Zhang, Yue Zhao, ShengHua Zhou, ZhiTao Wang
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930S (2023) https://doi.org/10.1117/12.2671305
In the satellite pose estimation problem, the deep learning method is used to train the network. The satellite pose needs to estimate the rotation (R) and translation (T), which are difficult to be well estimated simultaneously due to the internal coupling interaction. To solve the above problems, a dual-channel satellite pose estimation network based on ResNet50 is proposed to decouple the rotation and translation of satellite, effectively avoid the interaction, and estimate the translation and rotation of satellite respectively through the constructed network, which improves the recognition effect of satellite attitude. Through experimental verification, the network model constructed in this paper has better effect on the estimation of rotation and translation compared with other methods.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930T (2023) https://doi.org/10.1117/12.2671718
The ability of the separator to capture the context-detailed features of speech signals and the number of parameters directly affect the accuracy and efficiency of speech separation in time-domain speech separation network (TasNet). This paper combines lightweight external attention with convolution and extends external attention to channel dimension; while satisfying the fine-grained extraction and modeling of spatial-channel correlation, it maintains small parameters and computation. Convolutional position coding is also used to integrate the contextual relationship and relative position information of speech features better. The above module then applies as a separator in the encoder-decoder structure based on TasNet, and a new convolution-augment external attention model for time-domain speech separation is proposed: ExConNet. The comparative experimental results show that ExConNet achieves considerable accuracy of speech separation, while its model parameters and calculation amount are significantly reduced, which can better meet the need for efficiency of speech separation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930U (2023) https://doi.org/10.1117/12.2671925
The third-party emergency forces such as neighboring enterprises in the accident can improve the rescue efficiency and shorten the rescue time to ensure the production and property safety of the affected people as much as possible. This paper proposes an emergency material scheduling model taking time as the objective and considering the participation of multi-party emergency forces. The model can freely decide the type of materials to improve the rescue efficiency when the risk index of the accident point is satisfied. A real case illustrates the applicability of the proposed model. The result shows that the average arrival time of emergency supplies for corporate safety production accidents is 12.53 minutes, of which 89.13% of accidents can deliver supplies within 10 minutes. Daishan Fire Brigade and CNOOC (Zhoushan) are the main rescue partners for enterprises. In addition, foam trucks are given priority to be put into the rescue process. This study can provide a basis for enterprises to adjust the storage of materials and the government to supervise enterprises.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
JiuYuan Huo, GuanXiang Pei, TingJuan Wang, JinQuan Liu
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930V (2023) https://doi.org/10.1117/12.2671236
In view of the problems of BIM model, low data transmission efficiency, cumbersome data conversion and poor BIM rendering, An optimized and improved method is proposed to realize the lightweight of electric &electronic systems’ BIM model. This method uses RevitAPI to generate a custom GLTF model, split and preprocess the overall model, use the optimized model simplification method to ensure the details of the model, generate the custom GLTF data format, complete the lightweight of BIM model. Taking the model in the electric &electronic systems’ project as an example, a high-speed railway station and components were selected for this experiment to verify and compare the proposed method. The results show that the lightweight method of BIM model effectively reduces the model data quantity, improves the real-time rendering efficiency and optimizes the loading effect.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930W (2023) https://doi.org/10.1117/12.2672756
This paper analyses application status and technology of C-V2X intelligent sweeper, and research the test and evaluation method of C-V2X intelligent sweeper. Then through the domestic driverless sweeper test verification the test results meet the requirements of the test standard. This paper improves the test and evaluation system of domestic unmanned sweeper and provides reference for the test of domestic unmanned sweeper.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Dong Zhou, Feifei Liu, Xiangfei Dou, Jie Chen, Zhexin Wen
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930X (2023) https://doi.org/10.1117/12.2671480
At present, the detection of drainage pipe defects adopts manual frame-by-frame naked eye discrimination, which has low detection efficiency and high cost, so a two-path multi-receptive convolutional neural network is designed, which also takes into account a certain small volume on the basis of obtaining the highest classification index. The experimental results show that the volume accuracy of the designed model is 92.3%, the recall rate is 91.1%, the F1 score is 91.7%, the model volume is 30.7M, the parameter quantity is 8.97M, and the calculation amount is 2.25G. Compared with other networks, this model is more suitable for automatic identification of drainage pipes.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930Y (2023) https://doi.org/10.1117/12.2671812
Route planning is an essential and important part of unmanned aerial vehicle (UAV) operations at sea. Therefore, this paper designs the trajectory planning for autonomous obstacle avoidance of unmanned ships in complex environments. Adopt the body coordinate system and inertial coordinate system to confirm the coordinates and heading angle of the unmanned ship; improve the inertia weight, determine the space constraints of the track planning, and accurately determine the autonomous obstacle avoidance path of the unmanned ship. Simulation experiments show that the trajectory planning method for autonomous obstacle avoidance of unmanned ships in complex environments designed in this paper reduces the time consumption of navigation, has stronger real-time performance, and can approximately represent the global optimal trajectory.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Zhaolong Deng, Yanliang Qiu, Xintao Xie, Zuanhui Lin
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125930Z (2023) https://doi.org/10.1117/12.2671350
Considering the problem of the inability to obtain accurate depth information in 3D pose estimation, this research attempts to use a depth camera to obtain accurate depth information to solve this problem and achieve good results. In the process of research, it is found that the general object detection and evaluation method is not accurate enough under the framework proposed in this paper, so this research proposes an evaluation method suitable for this framework. A standardizer is also designed to optimize the detection effect while achieving efficient tracking objects. Ultimately, inference time is reduced by 35%. The implementation of this research architecture is open-sourced at https://github.com/DumbZarro/BuddHand.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259310 (2023) https://doi.org/10.1117/12.2672136
Coverage path planning algorithms are widely used by many robots conducting work like floor sweeping, map generating and underwater searching. One of the practical methods is the Backtracking Spiral Algorithm (BSA), which is efficient and complete coverage guaranteed. In practice, however, autonomous robots are faced with the problem of energy constraints. The duration of the battery of a robot is usually limited. This paper presents an adaptive coverage path planning method based on BSA considering the energy constraints. The objective of this method is to minimise the total length of the path the robot travelled to recharge while guaranteeing complete coverage. This method uses an estimation strategy which would use the energy consumption of the last round as a conjecture of that of the next round and recharge when the remaining energy would probably be insufficient to complete the next round while the robot passes nearby the recharging station. A simulation result is then provided to serve as proof of this method.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259311 (2023) https://doi.org/10.1117/12.2671503
This paper focuses on the recognition and classification of driver's dangerous driving actions through Blazepose algorithm and st-gru network to ensure that drivers can drive safely during the driving process and keep drivers safe at all times. blazepose is a lightweight human posture estimation model using blazepsoe method to replace the openpose method in human skeletal keypoints to improve the speed and reduce the model size. The st-gru network is one of the best action recognition models based on human skeletal keypoints, which is better than most of the current action recognition models in terms of model size, accuracy and recall value. Therefore, this project uses the st-gru network to classify the extracted human skeletal keypoint.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259312 (2023) https://doi.org/10.1117/12.2671448
With the rapid development of science, VR technology has become more mature and perfect, and has been widely used in various fields. Include human motion scanning, so as to improve the efficiency of human key point identification and interaction. Based on this, this paper designs an interactive and key point recognition scheme of human dynamic gestures by means of theoretical analysis and experimental analysis. Through experimental verification, it can be found that the gesture recognition rate of this scheme is high, with an average of 93.5% and high accuracy, and it can be popularized in practice.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Jing Yang, Caizeng Ye, Bei Han, Jilin Qin, Lei Peng
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259313 (2023) https://doi.org/10.1117/12.2671372
Cross-age image generation technology is to generate cross-age face images on the basis of the original face image. The synthetic face image can show facial details such as skin, wrinkles and hair at a certain age. The technology can be widely used in film and television, animation, public safety and other fields. Cross-age face synthesis techniques can be divided into traditional cross-age face synthesis techniques and cross-age face synthesis techniques based on generative adversarial network models. With the continuous development of GAN, the technologies based on generative adversarial network models have made more progress and advantages in the field of face synthesis. The model in this paper, based on the generation of the adversarial network model, combines the advantages of the conditional autoencoder and the StyleGAN model, and innovates in the use of the feature contrasting device, which can generate HD face images consistent with the change logic across ages, and effectively avoid the emergence of problems such as organ deformation and identity inconsistency.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259314 (2023) https://doi.org/10.1117/12.2671568
In order to explore the problem of dance somersault gesture recognition, a kind of dance somersault gesture recognition based on multi-scale depth feature fusion is proposed. Methods Through the information recommendation of key technical problems and solutions based on multi-scale depth feature fusion, the research of dance somersault gesture recognition was explored. The research shows that the efficiency of dance somersault gesture recognition based on multi-scale depth feature fusion is about 4.6% higher than that of traditional methods. The acquisition of main video information has always been inclined to obtain video key frames. However, in the face of videos with strong continuity and low repetition between human posture sequences in various movements, only key frames can't represent all the effective information of the videos. Most algorithms excessively pursue the differences between action categories, thus ignoring the degree of "cohesion" between simple actions within actions.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259315 (2023) https://doi.org/10.1117/12.2672158
Pedestrian detection technology has high application value in various fields, and deep learning has become a key development direction in computer vision. Human object detection has also shifted from traditional detection algorithms to deep learning. Due to the influence of complex light and obstacles in the station, as well as the occlusions and size changes of passengers, the algorithm must be optimized for these complex scenes. This paper takes pedestrian detection technology as the goal, compares the methods based on human body parts recognition from the concepts and classification of artificial neural networks and deep learning, and profoundly discusses the convolutional neural network based on deep learning. Finally, pedestrian detection algorithms' problems and future trends are compared and discussed.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259316 (2023) https://doi.org/10.1117/12.2672155
Stroke is a leading cause of disability in adults. Notably, about 75% of stroke survivors have upper limb damage, which greatly reduces the quality of life of the patient after recovery. The current routine rehabilitation recommendation is repetitive functional training (exercise-based training) to promote nervous system recovery, and then realize exercise rehabilitation. The cost, efficiency and success rate of traditional treatment methods are unstable due to various factors such as the professional level of therapists, the time required and the workload of therapists. In the case, rehabilitation robot-assisted therapy brings a new direction for the rehabilitation of stroke hemiplegia. In this paper, a new type of hand rehabilitation robot is designed based on the physiological structure of fingers, which is used to assist stroke patients in different stages of finger movement rehabilitation training. It can help the patient to practice grasp adduction and abduction repeatedly, reducing the burden on the patient. Secondly, in this paper, the degrees of freedom and movement of each finger joint are analyzed and calculated. Through modelling and finite element analysis based on Solid works to simulate the stress changes of exoskeleton in different rehabilitation stages, a model suitable for different stages of rehabilitation training is put forward.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259317 (2023) https://doi.org/10.1117/12.2671313
Microbial fermentation is a typical microbial fermentation process. Microbial bacteria ingest the nutrients of raw materials in the fermentation tank. Under appropriate conditions, enzymes in the body catalyze complex biochemical reactions to produce microorganisms. In order to guarantee the quality of modeling data and meet the accuracy, integrity, and consistency of data quality requirements, it needs to preprocess the input and output data. In this paper, the parameter model is solved by the particle swarm algorithm. Updating the parameter value of the next moment in real time constitutes a feedback correction to the prediction model. Theil inequality approach is adopted to test the tracking performance of the above model’s adaptive correction method. The Monte Carlo method is applied to generate multiple groups of different kinetic model values, which are substituted into the fermentation kinetic model as the real model parameter values. After the experimental analysis, the measured value of the model established by the method in this paper is closer to the predicted value, which has the effect of feedback correction and optimal control. The external conditions in the fermentation process are optimally controlled to achieve the effects of shortening the production period. It improves the yield of fermentation terminal target products and reduces the consumption of raw materials.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259318 (2023) https://doi.org/10.1117/12.2671555
Wearing a safety helmet can effectively reduce or prevent injury to the worker's head caused by hazardous materials in the construction site. However, due to poor supervision, safety accidents often occur when workers don't wear safety helmets. In this paper, we propose a safety helmet detection algorithm based on face detection and ridge regression. Firstly, we get the location information of the face box and the five key points of the face through face detection algorithm, and then get the helmet detection box corresponding to face through ridge regression model. We collected 4000 images of people wearing helmets for training and testing of ridge regression models. Compared with some of the most advanced methods, we have achieved very good results in the test set. The results show that mIoU reaches 70.118% and the detection rate is improved.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 1259319 (2023) https://doi.org/10.1117/12.2671681
To improve the efficiency and stability of power transaction management, a micro-grid power transaction management platform is studied. Block chain Internet technology is used to build the basic architecture of the electric energy transaction management platform. Based on the principle of low cost and high efficiency, the mathematical model of electric energy transaction is established. Combined with the distributed consensus mechanism of block chain technology, the intelligent contract of electric energy transaction is designed to realize the execution and management functions of electric energy block chain transaction. The test results show that the average time delay of the platform is about 0. 36 s for 10 groups of the random electricity trading instructions, which is real-time and efficient. For 5 random groups of users, the average electricity cost of the transaction results is 20. 42 yuan, which is feasible and economical, and lays a good foundation for the sustainable development of the micro-grid electricity trading management platform.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931A (2023) https://doi.org/10.1117/12.2671573
This paper constructs a 5G message-based smart energy efficiency billing system to address the data interaction and diversity problems in the digital transformation process in China's electric energy efficiency field. The publish/subscribe algorithm based on a hierarchical mechanism ensures high efficiency and stability in the process of data transmission. Through this smart energy efficiency billing system, the business service quality of State Grid Corporation of China can be improved to a great extent, and the efficiency of capital flow recovery of State Grid Corporation of China can be improved. The system provides users with convenient and comfortable services.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931B (2023) https://doi.org/10.1117/12.2671663
In view of the current TCP congestion control slow-start algorithm and its waste of bandwidth due to short connections, network congestion and packet loss caused by the rapid growth of the congestion window in the later period, this paper studies the slow-start algorithm part of the TCP transmission protocol. Considering the characteristics of the current relatively high-speed network, this paper proposes an improved slow start algorithm with traffic awareness. By statistical analysis of data transmission in the network, the algorithm dynamically determines the initial congestion window size of slow start, and dynamically adjusts the congestion window by tracking the changes of real-time network traffic. In the slow start stage, the smoothness of the congestion window is further analyzed, and the smoothness of the window growth is corrected in real time, so that the congestion window does not increase exponentially, but increases by a more efficient power function. The results of this experiment show that the improved algorithm slows down the growth rate of the congestion window and improves the smoothness of the window growth. It also significantly improved the data transmission rate and throughput.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931C (2023) https://doi.org/10.1117/12.2671300
Different from source coding, which only emphasizes coding efficiency, fault-tolerant coding adds some redundant information during coding to strengthen the ability of error resistance, so as to obtain the best gain with the least redundancy. Therefore, the application of fault-tolerant coding in the repair of mobile cache system of satellite Internet of Things can effectively improve the security of information storage of satellite Internet of Things. Based on this, this paper introduces the repair method of mobile cache system of satellite Internet of Things based on fault-tolerant coding from the aspects of data deployment and data collection, so as to repair the invalid data in time and ensure the integrity and availability of the saved data. At the same time, a specific method to optimize its system parameters is proposed. The repair of mobile cache system of satellite Internet of Things based on fault-tolerant coding can effectively solve the disadvantages of low price and short service time of sensor node terminal equipment in Internet of Things, and avoid problems such as data storage failure or performance weakening of cache communication system caused by sensor leaving satellite coverage and power exhaustion, so as to provide corresponding reference for the development of satellite Internet of Things.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931D (2023) https://doi.org/10.1117/12.2671494
Credit business income is the main source of income for banks, and effective prevention of credit risk is an important task for banks' operation and management. The application of various intelligent technologies in the financial field can provide strong technical support to the management of credit risk. How to use big data technology and artificial intelligence algorithms to improve risk control is an important research topic for commercial banks. To address the above issues, this paper studies the theories and technology applications related to artificial intelligence, risk management, and neural networks In this paper, by constructing a BP neural network model, determining evaluation indicators, and using model simulation and validation, risk assessment is performed on bank customer credit risk indicator data, and the validation reflects that the model has a better ability and high accuracy for customer risk prediction, which provides a reasonable determination of customer credit indicators and reduces It provides data basis for reasonable determination of customer credit indicators and reduction of bad debt losses of banks.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931E (2023) https://doi.org/10.1117/12.2671561
Features extracted by the neural network do not have scale invariance, which makes multi-scale image recognition and classification a difficult problem. Recent studies have proposed many new ways to solve this problem, such as feature fusion, sensor field transformation, etc. However, none of them essentially solve the problem that the neural network does not have scale invariance. In this paper, we propose a network generating network (NGN) architecture and design the NGNResNet network, which is an improved version of the ResNet network. The network can identify images at three scales simultaneously and has scale invariance. The experimental results show that the NGN structure helps us to improve the classification accuracy of small-scale images by about 10 percentage points, and helps to improve the performance of the network in the face of small targets.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931F (2023) https://doi.org/10.1117/12.2671289
Aiming at the complex and changeable driving scenarios of intelligent vehicles and the need to quickly and accurately identify obstacles, an improved YOLOV4 algorithm is proposed. To limit the number of neural network parameters, the CSP-darknet53 backbone of the original YOLOV4 was replaced with the Ghostnet backbone. In addition, to improve the neural network's accuracy, a lightweight attention mechanism ECA is added to the three effective feature layers generated by the backbone using residual block connections. Experiments have shown that the improved YOLOV4 has a 2.8% increase in mAP compared to the original YOLOV4. Without changing the accuracy, The network model's memory size is lowered by 39%, as well as a 50% improvement in detecting speed. Therefore, the improved YOLOV4 accuracy and real-time performance are better than the original network detection, providing a strong guarantee for intelligent vehicle obstacle avoidance.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931G (2023) https://doi.org/10.1117/12.2671306
China, with a large population, is a big producer and seller of agricultural and sideline products in the world, so how to sort them quickly and accurately is very important. In order to solve the problem of high-speed image acquisition and processing in color sorter, TMS320F2812 is used as the core processing chip, and the image processing algorithm is improved by the mixed mode of C language and assembly language. The color sorter control system based on DSP is designed by sorting execution module. It has been proved that the system can meet the design requirements of various performance indexes and has strong practicability.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931H (2023) https://doi.org/10.1117/12.2672765
The power network's operation security contributes to the power grid's smooth operation. Aiming at the problem that each network node in the conventional power communication network attack penetration system is fragile, and the global network attack graph cannot be generated, which leads to the failure of the network attack penetration vulnerability test, this study introduces the knowledge map into it and designs a power communication network attack penetration test system. In hardware, the FPGA chips and RAM are designed. In terms of software, the software architecture of the test system is established to control the network attack penetration globally, and then the knowledge map is used to construct the communication network attack graph model and generate the network global attack graph, so as to realize the effective test of the electric power communication network attack penetration vulnerability. By using the method of system testing, it is verified that the number of vulnerabilities tested by the system is consistent with the actual situation, and it can be applied to real life.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Ming Zhang, Xuan Yang, Zimu Wang, Lei Mao, Yini Zhao
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931I (2023) https://doi.org/10.1117/12.2672771
The capacitive voltage transformer operating condition monitoring method has the problem of excessive error, in order to design a digital twin-based capacitive voltage transformer operating condition monitoring method. The capacitive voltage transformer transmission characteristics are identified, the harmonic measurement signal is obtained by using a series-connected voltage divider, an equivalent circuit model is constructed based on digital twin, the capacitive transformer fault gas data is extracted and uploaded to the digital twin database, and the operating condition monitoring method is designed. The results show that the mean error value of this designed capacitive voltage transformer operating condition monitoring method is 24.334%, indicating that the capacitive voltage transformer operating condition monitoring method in the paper is more effective after combining digital twin technology.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931J (2023) https://doi.org/10.1117/12.2671278
Aiming at the positioning problems existing in rail transit system, this paper proposes an intelligent image recognition and positioning algorithm, which adopts deep learning technology to identify vehicles. And, through field test, the experimental results show the effectiveness of the algorithm. Its implementation cost is significantly lower than the existing equipment and can meet the requirements of the existing engineering practice.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931K (2023) https://doi.org/10.1117/12.2671264
Extracting road information from high-resolution remote sensing images is an important way to obtain basic data of geographic information. In this paper, firstly, the shortcomings of K-means and SVM are analyzed, and then the road information is extracted by the algorithm combining K-means and SVM. The experimental results show that the combined algorithm has higher accuracy and lower missing error than the single algorithm. The experimental results can provide some technical support for future road information extraction.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931L (2023) https://doi.org/10.1117/12.2672715
In order to accurately obtain the short-time OD passenger flow distribution of the subway line network, so as to efficiently coordinate the transportation capacity and passenger demand, a multi-time granularity subway line network short-time OD passenger flow prediction model based on LightGBM was constructed by combining the idea of ensemble learning. The model uses the subway automatic ticket sales and inspection data to analyze the temporal and spatial distribution characteristics of OD passenger flow on the line network, introduces a variety of temporal and spatial influencing factors to train and predict the data of the whole network, and studies the relationship between the prediction accuracy of the subway line network OD passenger flow and the time granularity. relationship between. Taking the Suzhou subway as an example, the results show that: compared with other models, the model can not only effectively reduce the prediction error, but also can effectively fit the peak passenger flow, and improve the accuracy of short-time OD passenger flow prediction of the subway network.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931M (2023) https://doi.org/10.1117/12.2671429
With the development of computer network technology, the risk of network intrusion also increases greatly. But the traditional encryption and firewall technology can not meet the security needs of today. Therefore, intrusion detection technology is a new dynamic security mechanism developed rapidly in recent years. This paper studies the security mechanism used to detect and prevent system intrusion. Different from the traditional security mechanism, intrusion detection has the characteristics of intelligent monitoring, real-time detection, dynamic response and so on. In a sense, intrusion detection technology is a reasonable complement to firewall technology.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931N (2023) https://doi.org/10.1117/12.2671871
Based on the BP neural network in the category of artificial intelligence technology, this paper combined with the conventional expert system diagnosis and discrimination method, and completed the construction of automatic fault diagnosis system for rotating parts of mining machinery in ASP.NET environment. Taking the common rotating machinery in mining machinery and equipment as the research object, aiming at the fault characteristics of mining machinery and the difficulties faced by maintenance, such as high difficulty, high cost and high risk factor, the system provides a new comprehensive application solution for the fault diagnosis of rotating machinery with the help of the application advantages of various information technologies. Through data feature extraction, automatic diagnosis, manual diagnosis, data management and other modules in the system, the whole life cycle management of mining machinery and equipment, early warning and treatment of faults, historical data query and other functions are realized. It not only improves the level of health management of mining machinery and equipment, but also establishes a solid guarantee for the safe and stable production of enterprises, and further makes a positive and beneficial attempt for the construction of smart mines in China.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Zhouzhou Wu, Weiwei Qin, Qiqi Lu, LongKun Wei, Long Wang
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931O (2023) https://doi.org/10.1117/12.2671841
There are many security risks in the data exchange of the Internet of Things. Therefore, this paper effectively integrates various technologies to design IoT security access gateway authentication to ensure the secure operation of IoT. The IoT gateway authentication hardware is divided into data collection, processing, access and power supply management modules to improve the security of IoT authentication and the efficiency of data processing. In the software design, the terminal device access identity authentication is realized through the communication protocol, and the identity authentication technology is analyzed to strengthen the security awareness of the network nodes. Test experiments show that in the face of different network problems, it can play a role in ensuring network security access and achieve more convenient, smarter, and more autonomous management.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931P (2023) https://doi.org/10.1117/12.2671661
Based on domestic cryptographic algorithms, this research encrypts the data of multi-system collaborative IoT terminal devices, and the data is transmitted in ciphertext to realize the security protection of massive structured and unstructured data; using database encryption and decryption, file system encryption and decryption and other passwords Technology to ensure data storage and data transmission security, to achieve data confidentiality, integrity and availability.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931Q (2023) https://doi.org/10.1117/12.2672207
Image processing is one of the most popular technology these days and is widely used in industrial robots and other fields. In the process of working with industrial robots, it is often necessary to locate the target through image processing technology and then process it later. In the process of object positioning, operators not only need its position in the image, but also the distance from the lens, and even the direction in which it moves. Compared to other object positioning methods this paper provides an object identification method that can find objects in the picture that are different from the background color precisely and calculate the distance from the camera and the direction of movement of the object. The method used black rectangle geometry on a white background as probe objects, and mainly used python based Opencv's algorithm for image processing. The experiment finally confirmed that the location, distance, and direction of motion of the target geometry can be well determined by this method.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Yanwei Wu, Yi Gao, Yingcheng Liu, Yangyu Zhao, Zijun He, Miaoyi Li
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931R (2023) https://doi.org/10.1117/12.2671915
In the valve hall of converter station, the operating environment has very strict requirements for the valve hall. Due to the space, equipment and other factors of the valve hall of converter station, it is very difficult for traditional monitoring methods to achieve the operating environment of the valve hall of converter station. In order to achieve online real-time monitoring, this paper adopts the method of multi-sensor distribution to achieve, simulates the converter station valve hall through modeling, and conducts reasonable sensor distribution to achieve all-round real-time monitoring of the valve hall, including temperature and humidity, electromagnetic strength, particles, etc. Through real-time monitoring, the measures can be pretreated to avoid accidents.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931S (2023) https://doi.org/10.1117/12.2671353
With the development of the Internet of Things, the world is entering an era of interconnection. Some typical application scenarios, such as environmental monitoring, energy management, space equipment operation and maintenance, require gateways to integrate multiple heterogeneous networks, such as WiFi, Bluetooth, Zigbee, LoRa and other wireless LAN and wired LAN. However, the interfaces of existing gateways are different and incompatible with each other, which makes difficult to achieve the requirements of heterogeneous interconnection. Therefore, this paper presents a smart integrated access gateway with modular architecture, which consists of a motherboard and multi-type user cards with pluggable functions. Different user cards can provide different communication interfaces and adapt corresponding communication protocols, by which different network customizations can be achieved in combination. Compared with other research work, the gateway is more configurable customizable and flexible.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931T (2023) https://doi.org/10.1117/12.2671334
The Internet of Things (IoT) and Machine Learning (ML) are two very hot technologies these days. IoT requires a lot of data processing, and ML is a useful means of processing data. Therefore, the combination of IoT and ML has become a very promising research direction. This paper is a investigation of the combination of IoT and ML. It first introduces the development history of IoT and ML, then introduces some achievements that have emerged in the field of ML and IoT combination. After that, the paper refers some ML technologies which will play important roles in IoT. In this process, this paper also proposes a scheme to improve the accuracy of YOLO algorithm by identifying picture groups. Finally, the paper discusses the existing problems and future development directions of the combination of IoT and ML and provides some references and suggestions for scholars who study the combination of ML and IoT technology.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931U (2023) https://doi.org/10.1117/12.2672131
The video surveillance command and management system based on GIS technology is studied, which enables users to interact with real scenes, and can effectively solve the spatial difference in multi-point surveillance. The system consists of two parts: hardware and software design. The hardware design includes intelligent monitoring front-end , transmission equipment and background monitoring center; the software design consists of GIS visualization display, scene fusion simulation and stereoscopic display. Through the test of the system, the demand for three-dimensional display of command and management information of the 3D GIS intelligent video surveillance system integrated with multiple scenes has been realized.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Shi Li, Lingling Han, Ke Huang, Fangzhen Ge, Lele Zhang
Proceedings Volume Second Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence and Big Data Forum (AIBDF 2022), 125931V (2023) https://doi.org/10.1117/12.2671844
In order to draw and analyze the current status of the research on the application of artificial intelligence in the field of health in China and abroad, we explored the hotspots and frontiers of this research field and the evolution path from 2012 to 2022. This research uses 3743 articles included in the core database of Web of Science and 728 related research articles contained in CNKI as the basis for data analysis, and uses the advantages of bibliometrics visualization software of CiteSpace and VOSviewer to study institutions, countries, authors and literature keywords. In this research area, the research of medical artificial intelligence is growing day by day. By analyzing and sorting out the current research status, research hotspots and evolution path in this field of research, and trying to point out the deficiencies in the research in this field in China, the paper hopes to provide more references for researchers and practitioners of medical artificial intelligence related research.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.