266results about How to "Accurate extraction" patented technology

The invention discloses a knowledge question-answering method and system based on a large language model and a semantic abstract. The system comprises a tree structure abstract generation subsystem which comprises a knowledge base configuration module, a file management module, a document analysis module and a knowledge block management module and is used for automatically generating tree structure hierarchical knowledge blocks for large manual documents or multi-chapter manual documents; the dual-channel retrieval engine subsystem comprises a question rewriting module, a semantic abstract retrieval module, a detail knowledge block retrieval module, a father-son retrieval module and a prompt word construction module, and all the modules work cooperatively to ensure that semantic abstract knowledge blocks and detail knowledge blocks related to the question of the user are retrieved. According to the method, by introducing the semantic abstract based on the tree structure and the two-channel retrieval, perception of global knowledge and accurate extraction of local fine-grained knowledge can be achieved at the same time, and then the ability of a knowledge base question-answering system in processing the general and inductive problems is improved.

The embodiment of the invention provides a modeling method and device for an airborne system function architecture model, computer equipment and a medium, and relates to the technical field of aviation architecture design, the method comprises the following steps: inputting a design requirement document into an airborne large language model, and in the large language model, under the structural constraint of an airborne system knowledge graph, establishing a design requirement document; the method comprises the following steps: extracting a logical relationship between entities required by airborne system modeling from a design demand document, screening a corresponding airborne meta-model according to the extracted logical relationship between the entities, and constructing a demand graph; constructing and outputting a text model of a concept scene, a text model of a task scene, a text model of a system function, a text model of a logic architecture and a text model of a physical architecture of the airborne system based on the demand graph; and inputting each text model into a system modeling language modeling terminal for visualization to obtain a multi-level functional architecture model of the airborne system. According to the scheme, the modeling efficiency, consistency and normalization are improved.

This invention discloses a primer and probe set, a kit, and their applications for detecting Staphylococcus aureus L based on real-time fluorescence MIRA technology. The primer and probe set consists of single-stranded DNA sequences 1-3. Based on the aforementioned primer and probe set, this invention establishes and optimizes a real-time fluorescence MIRA detection method for Staphylococcus aureus L. Furthermore, the optimized real-time fluorescence MIRA method is innovatively combined with a simplified nucleic acid extraction method and a portable fluorescence detection device, significantly reducing extraction and time costs while enabling rapid on-site detection of the virus, making it particularly suitable for grassroots units and remote areas. In addition, the real-time fluorescence MIRA detection method for Staphylococcus aureus L established in this invention has advantages such as high specificity, high sensitivity, and simple operation. Moreover, the detection process does not require opening the lid, effectively avoiding aerosol contamination, providing an effective technical means for the differential diagnosis and rapid detection of Staphylococcus aureus L.

The application discloses a data crawling analysis method and device based on an industrial internet, and a medium, and belongs to the technical field of industrial internet data processing, and is used for solving the technical problems that the industrial internet data crawling in the prior art cannot be compatible with multiple types of web pages and dynamic loading technology, lacks customized rule configuration, has weak anti-crawling capability, and has low data structuring degree. The method comprises the following steps: collecting user demand data of a user; performing customized configuration on the initial crawling rule under the relevant rule adjustment to obtain a customized rule; performing crawling processing on a target website under relevant dynamic web pages and static web pages, and based on an anti-crawling mechanism, monitoring and controlling corresponding crawling actions to obtain target webpage data crawled; performing data structure analysis processing on the target webpage data to obtain a data structure format; performing distributed storage on the data structure format to obtain data storage result information; and feeding back and adjusting the customized rule.

The invention discloses a power grid lightning arrester on-line monitoring system based on a PHM, and the system comprises a voltage on-line collection module which is used for collecting a voltage signal of a corresponding lightning arrester; the synchronous control module is used for acquiring synchronous sampling data of the leakage current signal and the bus voltage signal; the signal processing module is used for acquiring fundamental components of the leakage current signal and the bus voltage signal and harmonic components of preset times; the feature generation module is used for generating resistive current feature data of the lightning arrester; the data integration module is used for forming a multi-parameter operation state data set of the lightning arrester; the health state evaluation module is used for obtaining a health state evaluation result; the trend analysis module is used for acquiring lightning arrester health state change trend information; and the fault risk assessment module is used for generating a lightning arrester fault risk assessment result and realizing online assessment, trend early warning and fault risk prediction of the health state of the lightning arrester.

The invention provides a laser radar three-dimensional target detection method based on a coarse-to-mental network, and the method comprises the steps: firstly constructing a coarse positioning network based on a Transform network structure, and enabling the coarse positioning network to be used for positioning an interval to which an effective echo pulse signal belongs in one-dimensional time; and a fine positioning network based on a coding-decoding structure is further designed, and pulse peak positioning with higher time resolution is carried out on a high signal-to-noise ratio signal interval extracted by the coarse positioning network. The coarse positioning network of the method can eliminate noise signal segments irrelevant to effective echo pulses to the maximum extent, so that the influence of noise on the fine positioning network is reduced, and the signal-to-noise ratio of echo data and the three-dimensional target detection precision of the system are improved.

The invention discloses a judgment document element extraction method based on a large model, a storage medium and equipment, and the method comprises the steps: collecting a large number of judgment documents, carrying out the cleaning and word segmentation of each judgment document, and marking key element information; taking the judgment document as the input of the large language model, outputting a key element extraction result of the judgment document, setting a reward function based on the marked key element information, and performing feedback optimization on the large language model in combination with a dominant actor-commentator algorithm; and inputting a newly obtained judgment document into the optimized large language model, and extracting key element information. The judgment document element information can be accurately extracted, the analysis efficiency and accuracy of the judgment document are effectively improved, and manual intervention is reduced.

The invention provides a face recognition model training method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a plurality of face image samples, and inputting each face image sample into a pre-trained teacher model and a to-be-trained student model to obtain teacher features, student features and prediction categories of each face image sample; determining feature distribution loss information based on the teacher features and the student features of each face image sample, and determining feature classification loss information based on the student features and the prediction categories of each face image sample; total loss information is determined based on the feature distribution loss information and the feature classification loss information; and based on the total loss information, training a to-be-trained student model until a preset training condition is met, and obtaining a final student model as a face recognition model. By adopting the method, the recognition accuracy of the face recognition model can be effectively improved on the premise of considering efficiency and computing resources.

The invention provides a non-contact life signal processing method in a strong clutter environment, and belongs to the field of data processing.The method comprises the steps that extreme points in a signal data sequence of a user and first and last signal data are recorded as to-be-corrected data; obtaining an amplitude adjustment parameter and a slope adjustment parameter of each piece of to-be-corrected data according to a numerical value difference and an acquisition time difference between each piece of to-be-corrected data and the to-be-corrected data on the left side and the right side of the to-be-corrected data; and combining the numerical value of the to-be-corrected data to obtain a corrected value of the to-be-corrected data, thereby obtaining upper and lower envelope lines. The invention aims to solve the problem that the signal data of a user is processed by an EMD (empirical mode decomposition) algorithm due to the fact that an upper envelope line and a lower envelope line which are obtained through extreme points in a signal data sequence are greatly influenced by noise because the noise in an environment where the user is located has a great influence on the signal data of the user. And the accuracy of the obtained heart rate signal and respiration signal is low.

This invention discloses a UAV inspection method and system for bridge bearings. The method includes: planning a strip-shaped flight path and constructing a coarse 3D point cloud model of the bridge; then planning a scanning flight path around the bridge's outer envelope to collect fine point clouds of bridge components; firstly, using a small amount of labeled data combined with pseudo-label generation and knowledge distillation to complete lightweight training of the PointNet model; then, strengthening structural association features through a contextual attention mechanism and generating virtual points using an occlusion region feature completion method to complete the occluded region and obtain a complete bearing point set; finally, extracting bearing geometric parameters by weighted noise point removal, iteratively fitting bounding boxes, and smoothing; planning a flight path based on the bearing's location to collect the bearing's surface image; then performing image cleaning and enhancement processing, and finally identifying surface defects using image recognition methods to generate an inspection report. Comprehensive technical support is provided.

This invention discloses a requirement tracking method, apparatus, device, and medium based on a large language model, relating to the field of artificial intelligence technology. The method includes: performing hierarchical structure tracking and nested table linearization on the document to be processed, and semantically segmenting it; submitting each text block in parallel to a large language model to extract structured requirement items according to preset fields; calculating the cosine similarity between requirement items using an embedding model to recall a candidate matching set; then using the large language model to perform semantic association matching and coverage determination to obtain association mapping data; using the large language model to perform multi-level semantic matching on the requirement items, constructing a multi-level tracking link that supports cross-level matching between non-adjacent documents; and parsing the test item result description text and mapping it to a completed state when the link ends and contains a test document. This invention can improve the automation and accuracy of requirement item extraction, association matching, coverage determination, and full-link tracking, and enhance the fault tolerance of the tracking link.

The application belongs to the technical field of intelligent service robots, and relates to a home maintenance robot with omni-directional movement and electromechanical cooperative posture control, which comprises an omni-directional chassis, a multi-degree-of-freedom mechanical arm module, a capsule type modular intelligent material supply cabin, an environment sensing and interaction head module, a sensor group and a main controller. The chassis is driven by Mecanum wheels and motors; the material supply cabin comprises an annular cabin body assembly, a plurality of capsule cabins, a central feeding channel, a lifting feeding platform, a platform lifting driving mechanism, a cabin door opening mechanism, a discharging channel and a magnetic guide mechanism. The device can realize standardized classified storage of small standard parts, automatic cabin selection, single capsule release, central receiving and magnetic discharging, thereby reducing the risk of standard part scattering, mixing and jamming; effectively improving the passability of the robot in unstructured terrains such as carpets, doorsteps and slopes, avoiding the imbalance of the center of gravity when the mechanical arm extends for operation, and being compact in structure and suitable for daily maintenance of household items in single-family homes.

The invention discloses an RPA-based power dispatching service auxiliary method, system and device, and a medium. The method comprises the following steps: logging in a power dispatching OMS system by using an RPA automation technology; interactively triggering an operation management function through a target control, and sequentially executing UI element clicking, interface component activation operation and event distribution to a CLOW node to jump to a scheduling operation page; obtaining the number of executed work orders by adopting a selenium technology, triggering work order operation instructions in sequence, and extracting planned operation time, operation content and advance notice unit names in the work orders; on the basis of the extracted work order information, target grouping check and message content input and sending including operation content, time and ticket changing signature are completed in a message management module; generating massive controlled training sample sets through an RPA model, and performing parallel training on the sample sets by adopting a deep reinforcement learning algorithm; and carrying out adaptive process design based on a training result, and adjusting and perfecting a business process through iterative development testing and according to a deployment environment.

This application discloses an image feature extraction method, device, and storage medium. The image feature extraction method includes: extracting frequency domain information of an image to be processed; dividing the image to be processed into regions in the spatial domain based on the frequency domain information to obtain low-frequency image regions and other frequency domain image regions; removing interfering image information from the other frequency domain image regions to obtain frequency domain image templates corresponding to the other frequency domain image regions; performing attention calculation on the other frequency domain image regions and the low-frequency image regions based on the frequency domain image templates to obtain attention features corresponding to the other frequency domain image regions; and fusing the attention features and the image features corresponding to the low-frequency image regions to obtain the image feature extraction result. A morphological attention mechanism can be embedded into a neural network model. When the frequency domain features change, the morphological attention will also adapt accordingly, enabling joint optimization of model parameters and morphological attention processing to accurately extract image features.

This application discloses a method for inverting insect wingbeat frequency based on initial phase calculation, relating to the field of radar measurement technology. The method includes: establishing a radar echo model of insect wingbeat motion containing amplitude and phase initial phases; acquiring micro-motion signals from the collected insect echoes based on the model to generate a micro-Doppler spectrum; calculating the phase initial phase value of the insect wingbeat based on the micro-Doppler spectrum; constructing a phase compensation factor based on the phase initial phase value to obtain a wingbeat parameter plane; detecting energy peaks in the wingbeat parameter plane based on the phase compensation factor; and determining the insect's wingbeat frequency based on the position of the energy peaks. This application, by introducing phase and amplitude initial phases, recreates the dual physical processes of wing movement and body micro-motion during insect wingbeat, making the radar echo model highly consistent with the actual signal generation mechanism, reducing errors, and improving anti-interference capabilities.

The application relates to the field of heterogeneous bus communication and discloses a heterogeneous bus fault self-adaptive retransmission decision method and system, which solves the problem that secondary congestion and semantic fault are caused by the incapability of existing static retransmission to perceive cross-domain states. The method comprises the following steps: performing polymorphic fault fusion to output a comprehensive fault judgment signal; realizing standard scalarization and consistency checking of heterogeneous parameters through cross-domain state synchronization and a mapping table; calculating a real-time communication quality index, using clustering decision to determine an adaptive retransmission interval, number and path; performing semantic disassembly on cross-protocol domain data to build a data semantic structure tree, generating a compensation retransmission instruction and interweaving the instruction with a time sequence to issue the instruction. The application realizes a closed loop from perception to semantic compensation, improves communication reliability and can be applied to communication system equipment manufacturing.

The application relates to the technical field of data processing, and discloses a multi-modal neural biological signal processing method and device, a server and a storage medium. The method comprises the following steps: acquiring a multi-modal neural biological signal to be processed, and preprocessing the multi-modal neural biological signal; inputting the preprocessed multi-modal neural biological signal into a deep learning model, extracting deep features of each kind of modal neural biological signal in the multi-modal neural biological signal based on the deep learning model; inputting the deep features into a feature fusion layer for feature fusion to obtain target fusion features; inputting the target fusion features into a regression layer through a full connection layer, performing vital sign prediction on the target fusion features by using the regression layer, and generating a biological vital sign prediction result. Through implementation of the technical solution, the deep features can be effectively captured, the features do not need to be artificially designed or selected, the effective prediction of the biological representation of the neural signal is realized, and the prediction accuracy of the biological representation is improved.

The utility model relates to the technical field of cosmetic packaging, in particular to a serum protein active water film vacuum fresh-keeping glass bottle packaging device. The device comprises a storage assembly and a pumping assembly, the pumping assembly is connected with the storage assembly and extracts objects stored in the storage assembly outwards, the pumping assembly is externally connected with a smearing assembly, sealing assemblies are installed on the storage assembly and the smearing assembly, and the sealing assemblies control opening and closing installation. The storage assembly is designed to be in vacuum sealing, contact between oxygen in air and a serum protein active water film is remarkably reduced, the oxidation risk is reduced, the shelf life of the product is prolonged, the pumping assembly can accurately control the dosage extracted each time, waste is avoided, meanwhile, it is ensured that the dosage used each time is consistent, and the service life of the product is prolonged. The pumping mechanism reduces the opportunity that fingers of a user make direct contact with the product, pollution risks are reduced, product sanitation is guaranteed, the external smearing assembly is designed to be suitable for direct use, intermediate operation steps are reduced, and the user can smear the product easily.

This application discloses a method, system, device, medium, and product for 3D anomaly detection of industrial products, relating to the industrial manufacturing field. The method includes: preprocessing 3D point cloud data and 2D RGB images of the industrial product; extracting 3D point cloud features based on the preprocessed 3D point cloud data using a 3D point cloud feature extraction model constructed with stacked Point Transformer V3 as its backbone; extracting 2D image features based on the preprocessed 2D RGB image data using an image feature extraction model; performing multimodal feature fusion on the 3D point cloud features and 2D image features; determining an anomaly score based on the fused features and a preset mean of the fused features; and determining the anomaly detection result based on the relationship between the anomaly score and a set threshold. This application can efficiently process 3D point clouds, accurately extract 2D features, and achieve collaborative multimodal fusion.

This invention relates to the field of energy management technology, and in particular to a method and apparatus for energy management and adaptive power consumption control of unmanned systems. The method acquires multi-source operating data and real-time load location, extracts target energy operating data, collects dynamic power consumption fluctuation data for the next cycle based on the initial power consumption allocation scheme, and optimizes and generates an updated power consumption scheme based on the target data to achieve dynamic adaptation. Based on the current energy storage status, target data, and load priority, it calculates the extreme power supply range and determines the target power-protected load set to ensure priority power supply to critical loads. When it is determined that the system is about to run out of energy, it intelligently generates a target load-cutting sequence by comprehensively considering real-time operating conditions, priorities, and energy storage status, and executes orderly graded power outages, improving energy utilization efficiency. Under extreme operating conditions, it can scientifically plan power protection and load-cutting strategies, effectively avoiding disorderly power outages, and significantly enhancing the power supply reliability, safety, and intelligence level of unmanned energy systems.

This invention discloses a method and apparatus for fine extraction of aquatic vegetation and landscape connectivity assessment in rivers and lakes based on high-resolution remote sensing, belonging to the field of environmental remote sensing monitoring and water ecology assessment technology. It includes: acquiring and preprocessing high-resolution satellite imagery to obtain a standardized remote sensing image dataset; extracting water body masks and aquatic vegetation coverage areas through hierarchical threshold segmentation; constructing a dual-threshold discrimination rule based on the normalized vegetation index and soil-regulated vegetation index to finely divide the distribution areas of floating and submerged plants; constructing a binary raster map of vegetation distribution based on the distribution areas; combining a preset connectivity distance threshold; calculating the landscape connectivity index of vegetation patches based on graph theory; determining the ratio of equivalent connected area to effective connected area; and completing a comprehensive assessment of the ecological effectiveness of the aquatic vegetation community. This invention is mainly used for remote sensing monitoring of aquatic vegetation in urban rivers and lakes and for assessing the effectiveness of water ecology restoration, effectively mitigating spectral interference from complex water body sediment backgrounds.

The invention discloses an unstructured data report filling method and system based on a large language model, and belongs to the technical field of artificial intelligence, and the method comprises the steps: preprocessing unstructured data, and generating preprocessed unstructured text data; based on the preprocessed unstructured text data, constructing and executing a semantic structured extraction chain; field mapping and report template binding are carried out based on the constructed semantic structured extraction chain, and a field mapping rule is established; based on a field mapping rule, automatically generating and filling a report, and automatically generating the report according to the extracted structured data; and interpretable proofreading and manual auditing are carried out on the automatically generated report. According to the method, implicit information can be accurately extracted, and semantic inference and association are supported; automatic report filling is achieved, and the manual arrangement workload is remarkably reduced; and the output is explainable, and each field can be positioned to the original text, so that the auditing credibility is improved.

The invention provides a color feature extraction and correction method and system for drug sensitive plate MIC value interpretation and a computer readable storage medium. The method comprises the following steps: acquiring a to-be-processed drug sensitive plate image; performing visual preprocessing such as exposure, contrast and white balance on the image according to the configuration file; high-precision alignment correction of the image is realized by using ORB feature detection and a BFMatcher matching algorithm; generating a transparent mask with an Alpha channel based on predefined hole position contour data, and performing non-destructive target segmentation on the aligned image; and converting the target region to an HSV color space, extracting H-channel hue features, executing cyclic correction, eliminating the problem of numerical value jump of the red region, and obtaining stable color feature data. The method has the advantages of being high in image registration precision, accurate in segmentation, stable in color feature, high in robustness and the like, and automatic and standardized interpretation of the MIC value of the drug sensitive plate can be achieved under different shooting conditions.

The invention discloses a cultivated land non-agrochemical intelligent monitoring method and system based on multi-temporal SAR and optical image fusion, and belongs to the technical field of remote sensing monitoring and land resource management. The method solves the problems of information loss and misjudgment caused by cloud and mist shielding or noise interference of a single remote sensing data source. According to the scheme, the method comprises the following steps: acquiring and preprocessing a multi-temporal SAR and an optical image sequence; carrying out space registration; constructing a double-branch deep neural network, respectively extracting radar scattering features and spectral texture features, and fusing the radar scattering features and the spectral texture features by adopting a cross attention mechanism; generating a change intensity graph based on a time sequence change detection module; and extracting non-agrochemical suspected pattern spots through threshold segmentation, and aggregating and outputting a final monitoring result. The method is mainly used for realizing the automatic and high-precision monitoring of the non-agrochemical change of the cultivated land.

The application relates to the technical field of collaborative allocation of warehouse resources, and discloses a warehouse resource adaptive collaborative allocation system based on multi-time scale deep reinforcement learning, which comprises a warehouse data acquisition unit, which acquires equipment operation data and environmental parameters and standardizes the same; a warehouse resource state feature extraction unit, which extracts resource state features to form abnormal features; a reinforcement learning state space construction unit, which analyzes the features to form state space construction parameters; a resource allocation strategy generation unit, which constructs a model to generate a strategy and divides influence factors; and a strategy adaptive optimization unit, which updates and optimizes the strategy by using an algorithm. In addition, the system is provided with a multi-source information fusion unit and a task queue dynamic adjustment module. The system realizes adaptive collaborative allocation of warehouse resources, improves warehouse operation efficiency, adapts to dynamic changes of a warehouse system, and meets the demand of modern warehouse intelligent operation.

The invention is suitable for the technical field of mechanical equipment state monitoring and fault diagnosis, and provides a mine main conveying chain equipment bearing fault embedded diagnosis system and method.The mine main conveying chain equipment bearing fault embedded diagnosis system comprises a tunable piezoelectric energy collecting device which is installed on a bearing seat or a carrier roller shaft and used for converting vibration mechanical energy of a bearing into electric energy, a voltage signal representing the vibration state of the carrier roller bearing is synchronously output; the embedded wireless sensor node is electrically connected with the tunable piezoelectric energy collecting device and is used for managing the collected energy and processing and wirelessly transmitting the voltage signal; and the fault diagnosis module receives the voltage signal and performs feature extraction and fault classification on the bearing state based on a convolutional neural network algorithm. According to the invention, sensing and energy supply integration, frequency adaptive matching, high-quality signal acquisition and lossless embedded installation are realized, and a reliable, long-term and self-energy-supply bearing state in-situ monitoring and intelligent diagnosis scheme is provided for rotating mechanical equipment in various mine main transportation chain equipment.

This invention discloses an early warning method and system for high and low voltage switchgear, belonging to the field of power system automation. The method includes: constructing a multi-dimensional sensing network to collect multi-source heterogeneous data such as temperature rise, partial discharge, mechanical characteristics, and environmental conditions in real time; preprocessing the data and aligning it with spatiotemporal features, and dynamically correcting the sensitivity threshold of the partial discharge criterion based on humidity or pressure change rate; extracting multi-dimensional fault feature vectors and dynamically assigning weights to each component according to operating conditions and environmental conditions; inputting the weighted feature vectors into a long short-term memory network integrating an attention mechanism to generate a multi-timescale predicted trajectory of the operating state; comparing the predicted trajectory with preset multi-level state thresholds in real time, and executing a three-level early warning judgment (attention, abnormal, and severe) and corresponding linkage protection strategies. This invention can reduce the risk of false alarms and missed alarms, providing reliable support for the intelligent operation and maintenance of high and low voltage switchgear.

This invention discloses a method and system for defect identification of molybdenum-rhenium alloy pipe fittings based on high-definition microscopic images. The method includes the following steps: acquiring high-definition images of the molybdenum-rhenium alloy pipe fittings using a high-definition microscope; preprocessing the high-definition images of the molybdenum-rhenium alloy pipe fittings to obtain an image of the molybdenum-rhenium alloy pipe fittings to be detected; performing region growing algorithm processing on the image of the molybdenum-rhenium alloy pipe fittings to be detected to obtain a defect region to be detected; identifying defects in the defect region to be detected to obtain the type of defect; and determining that the molybdenum-rhenium alloy pipe fittings has a defect problem based on the type of defect.

The application discloses a kind of based on acoustic wave FMCW's respiratory frequency detection method, belong to acoustic wave signal processing technical field;Method is: utilize mobile phone loudspeaker to generate a segment frequency continuous change acoustic wave signal, form continuous acoustic wave detection sequence;Mobile phone microphone receives the acoustic wave signal that is reflected back synchronously;To the acoustic wave signal that is reflected back received filtering processing;With reflected signal, the distance change curve of thoracic cavity with breathing movement is generated by autocorrelation operation to transmission signal;Distance change curve is carried out smoothing processing;The main frequency frequency is obtained by fast fourier transform, and respiratory frequency is calculated.The present application effectively suppresses environmental noise interference by using the advantages of FMCW technology, analyzes the autocorrelation coefficient of reflected signal, accurately extracts the thoracic movement information related to respiration, and realizes the accurate detection of human respiratory frequency;Using autocorrelation function is easier to find peak, more easily tracked nature, greatly increase the convenience of user.