2799 results about "Network on" patented technology

The invention provides a multi-platform cooperative unmanned inspection task planning method and system for a wind power scene, and relates to the technical field of wind power plant inspection, and the method comprises the steps: based on a wind power plant three-dimensional scene model, adaptively dividing an inspection region, constructing a reachability path network, calculating an inspection difficulty coefficient according to target detection point information, and calculating a target inspection point; and generating an unmanned aerial vehicle and unmanned vehicle inspection area. A sensor is deployed to collect environment and equipment state data, feature vectors are extracted to establish an inspection constraint model, and main and branch inspection paths are planned on the reachability path network. And performing multi-platform relative positioning by using ultra-wideband radio waves, and dynamically optimizing an inspection path according to the equipment health score and the inspection constraint model. Through multi-platform cooperation, adaptive path planning and intelligent state identification, the wind power plant inspection efficiency and accuracy are improved, and the inspection cost is reduced.

The invention discloses a pipe network on-line monitoring system, which relates to the field of urban capital construction and comprises a sensing deployment module, a signal acquisition module, a signal processing module, a pipe network modeling module, a leakage positioning module, an edge calculation module, an energy scheduling module, an anomaly prediction module, a deposition evaluation module, an early warning visual module and a center management module. According to the method, the recognition capability of the system on a hidden and complex pipeline structure is improved, accurate and efficient three-dimensional modeling can be carried out, the path deviation resistance capability is improved, and the more comprehensive and robust leakage sensing capability is realized; the method is advantaged in that communication burden and response time delay are substantially reduced, sensitivity and foresight of abnormal evolution trend identification are improved, overall identification accuracy of the system is improved, a global optimal risk solution can be efficiently output, a high-risk area is visually displayed, and rapid pre-judgment and decision making of operation and maintenance personnel are facilitated.

The invention discloses a multi-modal industrial data exception collaborative detection method and system. Multi-modal data in the industrial production process are collected in real time through a data collection module. The data preprocessing module is used for cleaning, integrating and standardizing the acquired data and eliminating noise and abnormal data; and the multi-modal data fusion module fuses the preprocessed data and constructs a multi-modal data fusion graph. And the anomaly detection module applies a graph attention mechanism network on the fusion graph, captures the complex mutual relation of different modes on the global sequence level in time and space, and performs anomaly detection by using a dichotomy model. The root cause positioning module performs root cause positioning in the fusion image by using a random walk algorithm after detecting an anomaly, and determines the first several abnormal root causes by comparing weights of different root causes. According to the system, high efficiency and high quality of industrial anomaly detection can be improved in various complex industrial application scenes.

The present invention relates to the technical field of railway inspection, and relates in particular to a railway foreign-object intrusion detection method and system based on deep learning. The method comprises: S10, acquiring image data to undergo detection; and S20, inputting the image data into a trained attention semantic segmentation network to obtain a foreign-object detection result. The attention semantic segmentation network is obtained by first training a preset attention semantic segmentation network architecture using a preconfigured data set, and then re-training the attention semantic segmentation network on specified image data using an adaptive correction algorithm. A backbone network architecture is obtained by inserting a specified attention mechanism at a specified position within a pre-selected residual neural network, and using three parallel dilated convolutions as an initial convolutional layer in the residual neural network. A dual-branch decoder combines an edge recognition branch and a semantic segmentation branch. The method is applicable to various scenarios, achieves improved detection accuracy and maintains a lightweight design.

The invention discloses an improved YOLOv8-based penicillin bottle body defect detection method, and belongs to the technical field of computer vision, and the method comprises the following steps: obtaining and preprocessing penicillin bottle body defect data to obtain a training set and a test set; establishing a defect target detection model of the improved YOLOv8; training a defect target detection model by adopting the training set, optimizing a loss function, and updating a model weight parameter until the loss function is converged; and testing the defect target detection model by adopting the test set. According to the invention, a space separable pooling attention module SSPA is introduced, so that the perception capability of the network on subtle features such as bottle body scratches can be enhanced; a feature enhancement module MMFF is introduced, convolution kernels of different sizes are adopted to extract defect features of different sizes, such as smudginess, scratches and loading quantity anomaly, under a complex background, and the feature fusion effect is improved by module stacking.

The invention discloses an intelligent storage equipment health state assessment and predictive maintenance method and system, and the method comprises the steps: arranging a multi-mode sensor network on equipment, collecting vibration, stress, temperature, current, voltage and image data in real time, carrying out the collection and preprocessing of edge data, and transmitting the data to an analysis layer; and the analysis layer performs online self-learning calibration by using a digital twin model, fuses the data in combination with a fuzzy neural network for dynamic weight adjustment, and calculates a health index and a residual life prediction value of the equipment. The system can give out an early warning before a fault symptom occurs, and automatically generates a maintenance decision and a task scheduling scheme, thereby achieving the real-time monitoring and predictive maintenance of the equipment state, and remarkably improving the operation efficiency of a warehousing system and the reliability of the equipment.

The invention discloses an adaptive game-driven network defense method and system, and relates to the technical field of network security. The method comprises the following steps: generating a multi-modal bait according to network context information, and screening an optimal bait through credibility evaluation; constructing a time sequence feature tensor according to the network event sequence information of the bait, and obtaining predicted attack information by adopting a pre-trained attack prediction model; combining the network event sequence information of the bait and the predicted attack information to construct a defense income matrix, and carrying out iterative equilibrium solution to obtain an optimal strategy candidate pool; and taking the defense hybrid strategy of the optimal strategy candidate pool as an initial population, performing multi-objective optimization through a non-dominated sorting genetic algorithm to obtain a Pareto optimal strategy set, and performing screening to obtain an execution strategy set for dynamic defense decision making. According to the invention, the dynamic property, intelligence and self-adaptability of network defense are realized, and the ability of a network system to cope with complex attacks is effectively improved.

The present invention discloses an integrated circuit chip with an address code, a key, and a mechanical keyboard. The integrated circuit chip contains a programmable address code, integrates a magnetic induction circuit, a GRB magic color drive circuit, and includes a power pin, a ground pin, and a data pin. The power pins of n integrated circuit chips are connected in parallel to share a power network VCC, the ground pins are connected in parallel to share a ground network GND, and the data pins are connected in parallel to share a data line network DIO connected to an MCU, and data information is transmitted on the data line network DIO. Another integrated circuit chip with an address code only requires a power network VCC and a ground network GND, and data information is modulated and transmitted on the power network, which greatly saves wiring space, reduces the difficulty of PCB design, and exponentially reduces the number of electronic components. In addition, a variety of different customized code sending methods and working modes are provided, making its application simpler, more efficient, and more reliable.

The invention discloses a sparse view angle three-dimensional reconstruction method based on 3DGS. The method comprises the following steps: providing a Gaussian field initialization model; according to the method, the FPN, the 2D U-Net and the MLP are used for forming a Gaussian parameter initialization network, the network integrates high-level and low-level semantic information, and the understanding of the network on multi-scale features is enhanced. And then a complete Gaussian field model is obtained by combining initial point cloud coordinates generated by DUSt3R. And then performing projection and rasterization under different visual angles on the 3D Gaussian ball by taking three-dimensional Gaussian splashing as a main body frame of an algorithm, performing adaptive adjustment of cloning and pruning on the 3D Gaussian ball with relatively large gradient change, and outputting an output result which is optical information and depth information of each visual angle image after rendering. Then, in gradient back propagation, pixel value absolute value deviation between the RGB image obtained by rendering and the RGB true value is supervised through L1 norm loss; the scene rendering quality is high, calculation is simple, and engineering implementation is easy.

An efficient heterogeneous federated learning method based on hybrid distillation includes: initializing, by a server, global model parameters, and setting a preset total number of training rounds and a number of clients participating in each of the training rounds; loading local datasets in the clients respectively, performing random transformations on the local datasets to generate client distillation data for the clients, sampling multiple sub-networks from an original network of each client, training each sub-network on the client distillation data to obtain updated local model parameters of each client, and uploading the updated local model parameters to the server; and receiving, by the server, the updated local model parameters, performing, by the server, server distillation based on the updated local model parameters and a preset auxiliary dataset to obtain updated global model parameters and an updated global model, and sending, by the server, the updated global model to the clients.

Methods, systems, and non-transitory computer readable storage media are disclosed for modifying digital images via a generative neural network with local refinement. The disclosed system generates, utilizing an encoder neural network, a latent feature vector of a digital image by encoding global context information of the digital image into the latent feature vector. The disclosed system also determines a modified latent feature vector by trimming the latent feature vector to a feature subset corresponding to a masked portion of the digital image. Additionally, the disclosed system generates, utilizing a generative decoder neural network on the modified latent feature vector, digital image data corresponding to the masked portion of the digital image. The disclosed system also generates a modified digital image including the digital image data corresponding to the masked portion combined with additional portions of the digital image.

Embodiments are described herein for document authentication certification using information stored on a distributed ledger such as a blockchain. A distributed ledger may securely store document data describing the document. Use of a distributed ledger may provide an immutable, readily auditable record of the history of the document. Each user participating in the system may be assigned a unique identifier to be used for conducting transactions on the distributed ledger network. A user may also be provided with a digital security token such as a cryptographic key that is useable to authenticate the user and enable access to the document data stored on the distributed ledger(s).

The invention discloses a battery health state and residual life combined determination method, electronic equipment and a medium, and relates to the technical field of battery management. The method comprises the following steps: acquiring battery operation parameters at the current moment; constructing a hybrid model based on a physical model and a data driving model, and controlling the hybrid model through an adaptive gating mechanism to determine a current battery health state based on the battery operation parameters; the physical model is realized by introducing a thermal model on the basis of an equivalent circuit model, and the data driving model is realized by introducing a lightweight neural network on the basis of a neural network model; the current battery remaining life is determined based on the current battery health state through a life prediction model and an optimization algorithm, the life prediction model is achieved based on at least one of degradation curve extrapolation and multi-step rolling prediction, and the optimization algorithm is achieved based on Bayesian optimization or reinforcement learning. According to the scheme, the accuracy of determining the battery health state and the residual life can be improved.

The invention belongs to the technical field of power information security, and discloses a power network context scene association access control method and system, and the method comprises the steps: obtaining an access request and multi-dimensional context information associated with the access request, the multi-dimensional context information comprises at least one of time information, position information, equipment state information, user behavior information and network environment information, analyzing the multi-dimensional context information, identifying a power business scene to which the current access request belongs, and dynamically generating or adjusting an access control strategy for the current access request; and performing an access control operation on the access request based on the dynamically generated or adjusted access control policy. According to the method, the access control strategy can be dynamically adjusted according to the real-time context and the specific service scene of the power network, the accuracy, the flexibility and the safety of access control are improved, and the method effectively adapts to a complex and changeable power network operation environment.

The invention provides a smart power grid cloud edge collaborative heterogeneous data security access method, which comprises the following steps: collecting multi-source heterogeneous power equipment data in real time based on a deployed multi-mode sensor network; based on a lightweight federated learning edge inference engine deployed on an edge end network, performing adaptive semantic enhancement preprocessing on the multi-source heterogeneous power equipment data to generate an edge encryption feature tensor; performing three-dimensional credible security authentication and knowledge fusion on the edge encryption feature tensor based on an electric power space-time knowledge graph fusion engine deployed on the cloud side to generate secure and credible electric power data; and performing space-time causal reasoning and twin mirror image comparison on the secure and credible power data based on a federated block chain-driven digital twin decision engine deployed at a cloud end to generate a dynamic security policy map and store the dynamic security policy map. According to the scheme, the data security is enhanced, intelligent analysis and dynamic decision can be performed according to the real-time state of the power grid, and optimal scheduling and stable operation of the power grid are realized.

To provide a mechanism capable of easily searching for information indicating a file storage location without being restricted by circumstances under which the file has been stored on a network.SOLUTION: A user terminal 101, which can be connected via a network 105 to a large language model server 103 in which a large language model searches for URLs being information indicating file storage locations on the basis of similarity of language vectors of natural language, comprises a CPU 201 that transmits, in natural language, a query sentence for inquiring about a file storage location to the large language model, and receives, from the large language model, a response sentence that conveys a URL being information indicating the file storage location searched by the large language model using a language vector of the natural language of the query sentence.SELECTED DRAWING: Figure 1

Systems and methods for tokenizing entities within a distributed ledger technology (DLT) network. A processing system operating in the DLT network receives an input dataset that includes asset attributes, identity information, or event metadata. The processing system encrypts the input dataset, partitions the encrypted data into multiple encrypted segments, and distributes each segment to selected decentralized storage nodes. Each encrypted segment is associated with a unique segment identifier. The processing system generates a cryptographic token directly linked to the input dataset and embeds metadata into the cryptographic token referencing encrypted segments, a data provenance identifier, or a creation timestamp. A smart contract deployed on the DLT network mints the cryptographic token and cryptographically links the embedded metadata to encrypted segments stored at decentralized storage nodes. The processing system, through the smart contract, provides authorized access to the input dataset according to access permissions defined within the smart contract.

The invention discloses an electrocardiosignal classification method based on a multi-threshold pulse neural network, which comprises the following steps: acquiring an original electrocardiosignal, and denoising the original electrocardiosignal by using wavelet transform; segmenting the denoised electrocardiosignals into cardiac beat segments according to R wave crest positions, and dividing the cardiac beat segments into a training set and a test set; based on the discharge characteristics of the biological neurons, multi-threshold integral pulse distribution neurons are constructed, and a multi-threshold pulse neural network is constructed; training and testing are carried out, a gradient agent method is used for adjusting parameters of the pulse neural network, an optimal configuration network is obtained, and classification of electrocardiosignals is achieved. According to the method, the extraction capability of the pulse neural network on the electrocardiosignal time sequence characteristics in the training process is improved; and meanwhile, the original sparsity and biological inspiration characteristics are reserved, so that the method not only has excellent performance in the aspect of processing time sequence data, but also can remarkably reduce the calculation energy consumption, and is suitable for portable medical equipment and resource-limited edge equipment.

The invention discloses a static electromagnetic calculation method of a physical information neural network based on Fourier mapping, and the method comprises the steps: constructing a full-connection network structure fusing a Fourier mapping mechanism, and designing a parallel modeling architecture of a plurality of three-output sub-networks to adapt to multi-region medium division, network output predicts vector magnetic potential, normal magnetic field intensity and tangential magnetic field intensity at the same time, the problem of network training instability caused by medium mutation is relieved, and meanwhile the network training process has higher physical interpretation performance; vector magnetic potential and magnetic field component continuity are introduced as interface coupling physical constraints, and global physical consistency of solutions is ensured. A Fourier feature mapping mechanism is introduced into a physical information neural network structure, so that the expression capability of a neural network on high-frequency spatial features is enhanced, and efficient modeling of complex boundaries, medium distribution and multi-source coupling problems in a two-dimensional static electromagnetic field is realized; explicit annotation data is not needed, and the method is suitable for modeling and prediction of various static electromagnetic problems.

The invention discloses a multi-modal data fused shield tunneling machine cutter state fault prediction system, which comprises a state acquisition module for acquiring multi-modal state data, decomposing the multi-modal state data into time and space related characteristics, and obtaining optimal state data through low-rank matrix decomposition and compression; the anomaly simulation module is used for collecting fault state data, constructing an improved lightweight weight migration network to perform cross-domain feature extraction to obtain a cross-domain common feature vector, and obtaining common features through a feature alignment algorithm based on adversarial training on the basis of the cross-domain common feature vector and the fault state data; and the fault prediction module is used for performing fault prediction through a space-time cross attention dynamic migration network on the basis of the obtained optimal state data and common characteristics, so that long sequence characteristics can be captured when the cutter is abnormal, the calculation amount can be reduced when the cutter is normal, and the real-time performance and the accuracy of fault prediction are improved.

The present invention relates to the technical field of energy meter electrical parameter measurement. Disclosed is an intelligent anti-interference system for a power line carrier energy meter. The system comprises: collecting voltage and current signals of a power line, and constructing a sample database; performing time-frequency feature classification by means of a preset model, and establishing feature and scenario mapping by synchronizing scenario recognition results; building a scenario interference recognition network on the basis of the mapping result; collecting real-time data of N communication channels of a power line carrier energy meter, and performing channel coordination decision and anti-interference enhancement; performing communication data evaluation on the basis of a comprehensive scoring function; and performing anti-interference management on the basis of channel access decision, anti-interference enhancement, and a parallel access result. The present invention solves the technical problem of poor accuracy and reliability of metering data of existing power line carrier energy meters caused by interference signals and noise when the power line carrier energy meters are in complex electromagnetic environments, and achieves the technical effect of improving the anti-interference capability and the reliability of the power line carrier energy meters.

A system includes multiple network devices and one or more processors. The network devices are to connect to a network. The one or more processors are to exchange communication traffic over the network via the multiple network devices, to estimate multiple communication loads experienced respectively by the multiple network devices, and to distribute subsequent communication traffic among the multiple network devices, responsively to the multiple estimated communication loads.

The invention discloses a space-time adaptive power prediction system and method for distributed photovoltaic power generation in a mountainous area, and particularly relates to the technical field of fault prediction and health management. The method is used for solving the problem that the accuracy of power prediction and equipment health state evaluation is influenced by environment characteristic data reconstruction reference drift caused by equipment performance degradation in the prior art. A spatio-temporal data sequence is constructed by acquiring historical data of a photovoltaic unit, a spatio-temporal association network of generated power among equipment is constructed, performance stability of a reference equipment group is evaluated, and when the stability does not meet conditions, a performance attenuation spatio-temporal mode is analyzed to identify common features and personalized features. The propagation path of performance degradation on the space-time correlation network is analyzed based on the characteristics, a critical point is evaluated, reference correction is performed on the environment characteristic data by using the common characteristics and the critical point, and finally, the corrected environment characteristic data is used for executing power generation power prediction and equipment health state evaluation. Therefore, the prediction accuracy and the health management reliability are improved.

The invention provides a ship target detection method based on nonlinear network enhancement, a medium and equipment, and belongs to the field of artificial intelligence. The method comprises the following steps of: performing type conversion, data division and enhancement on image data in a data set by adopting the ship data set acquired in a real river channel scene; a nonlinear network enhanced YOLO ship detection model is constructed; adopting a multi-task joint loss function to train the YOLO ship detection model; and inputting a port monitoring image and a sea surface aerial image into the trained YOLO ship detection model, outputting a category number, a confidence value and bounding box coordinates of each prediction box, and forming a visual image. According to the method, a nonlinear network enhanced YOLO ship detection model is constructed, and the modeling expression capability of the network on a ship target in a complex scene is effectively improved, so that the robustness and the accuracy of target detection are enhanced, and particularly, the performance is better under the conditions of small target detection and image degradation.

The invention discloses a double-branch infrared small target recognition method based on a selective state space, and the method comprises the steps: collecting and dividing an infrared small target data set, carrying out the feature extraction of the infrared small target data through employing a backbone network Resnet-50, inputting a region-of-interest prediction module RPN and an attention aggregation module KAAM, and carrying out the recognition of an infrared small target. Semantic representation and position information are further extracted and fused, and accurate recognition of the small target under the complex background is achieved; and through a target discrimination network, introducing an SS2D spatial selective modeling module, deeply extracting global context information, and dynamically modeling a space-time dependency relationship between a target and a background. According to the method, the recognition limitation of a traditional method in the face of complex background interference and target size change is effectively overcome, meanwhile, an optimization strategy based on comparison constraint is designed in the training process of the network model, the recognition and segmentation precision of the network on a tiny target is enhanced, and high-robustness small target recognition performance is achieved.

Disclosed embodiments receive an event stream from a remote capture agent. The event stream includes timestamped event data generated by the remote capture agent based on network traffic monitored by the remote capture agent. A graphical user interface (GUI) is caused to be displayed for obtaining configuration information for configuring the generation of time-series event data from network packets captured by the remote capture agent. A set of statistics is generated from the time-series event data, and the configuration information is updated to trigger subsequent storage and processing of at least a portion of the event stream by one or more components on a network based on one or more of the statistics, a storage limit associated with the time-series event data, an index volume of the event stream, a historical trend associated with the statistics, or input through the GUI.

The invention discloses an emergency patient information data matching method and system based on the Internet, and the method comprises the steps: obtaining multi-source heterogeneous data in real time according to an electronic medical record, image data and real-time vital sign data of an emergency patient, and generating a time-synchronized multi-source data set through a self-adaptive data format conversion algorithm; for the multi-source data set, using a feature extraction network to extract illness state features of patients, and generating patient priority scores; adjusting a resource allocation scheme in real time by adopting a medical resource matching algorithm based on a graph neural network according to the patient priority score and the illness state feature, and generating a preliminary medical resource matching result; and for a medical resource matching result, generating a final emergency patient information data matching scheme through a differential privacy protection mechanism by adopting a data security transmission protocol based on homomorphic encryption. According to the embodiment of the invention, the method can meet the requirements of quick, safe and effective processing of data in an emergency scene, and improves the quality and efficiency of emergency medical service.

Embodiments herein provide methods for HPLMN control for discontinuous coverage in satellite access network by UE. The method includes determining whether DisCoNoserviceapplicability parameter is set to a value corresponding to true to indicate that UE remain in no service or whether parameter is set to a value corresponding to false to indicate that UE selects and camps on alternate network. In an embodiment, the method includes remaining in no service during discontinuous coverage duration and wait for completion of discontinuous coverage to resume normal services when DisCoNoserviceapplicability parameter is set to value corresponding to true to indicate that UE shall remain in no service. Alternately, the method includes performing PLMN selection procedure to camp on alternate network during discontinuous coverage to receive normal services when DisCoNoserviceapplicability parameter is set to the value correspond to false to indicate that UE selects and camps on alternate network.

Embodiments manage data in a file system over a network. A plurality of file system operations in the file system may be executed based on a file system client action or a file system administrative action such that the file system may be integrated with a network. Characteristics of the plurality network components in the network infrastructure that may be associated with the file system may be determined. Tests may be generated based on the characteristics of the network components such that the tests may be executed to evaluate the network components. Results of the tests may be employed to perform further actions, including determining non-compliant network components based on the results; modifying the network infrastructure based on the non-compliant network components such that one or more of file system operations are modified based on the non-compliant network devices; executing the modified file system operations on the modified network infrastructure.

The invention relates to the technical field of structural engineering, and discloses a post-fatigue residual stiffness calculation method for an assembly type composite beam, which comprises the following steps: acquiring three-dimensional laser point cloud data and thermal infrared image data of the composite beam, and registering the three-dimensional laser point cloud data and the thermal infrared image data; quantifying the macroscopic geometric defect based on the multi-modal data; welding area features are extracted, and micro-crack forms are recognized; constructing a microcrack network diagram and extracting parameters; evaluating the influence of the microcrack network on the local rigidity by adopting a self-attention map neural network, and obtaining a rigidity parameter after multi-scale damage correction in combination with the influence of the macroscopic geometric defect; calculating residual stiffness after fatigue by combining fatigue load history; according to the method, the initial geometric defect and the early microscopic damage can be comprehensively considered, and the accuracy of the residual stiffness calculation of the fabricated composite beam is improved.