1010results about How to "Improve targeting" patented technology

The invention discloses a design manuscript code generation and visual error iterative correction method and system based on multi-agent cooperation, and aims to solve the problems of low front-end code generation efficiency, poor visual restoration degree and lack of an automatic correction mechanism in the prior art. The system constructs a four-node state diagram based on LangGraph, automatically initializes a series engineering scaffold, obtains design data and generates a structured protocol, and generates a component code based on the protocol and performs visual verification and iterative correction on four processing nodes. The method comprises the following steps: dynamically generating an engineering scaffold through a large language model agent, converting Figma design data into a structured protocol, and carrying out iterative optimization by utilizing a judger-repairer mode until an error is lower than a threshold value; according to the method, high-fidelity automatic delivery from the design manuscript to the code is realized, and the executable front-end engineering code and the visual reduction degree analysis report are output.

The invention relates to the technical field of multi-beam sounding calibration, and discloses a multi-beam sounding online real-time calibration method. The method comprises the following steps: acquiring original water depth data and corresponding sensor attitude data in real time; generating a data quality evaluation result containing a data credibility index and a data abnormal distribution feature based on the original water depth data; triggering a dynamic calibration mechanism containing a calibration period adaptive adjustment strategy and a calibration parameter optimization rule according to the data credibility index; optimizing the target calibration rule parameters in combination with the sensor attitude data to generate an optimized calibration parameter set, and correcting the original water depth data in real time by using the parameter set to generate primary calibration data; and performing topographic feature consistency check on the primary calibration data, if the primary calibration data does not pass the topographic feature consistency check, starting a re-calibration process containing a rule base updating mechanism and a parameter iteration mechanism, and analyzing a deviation trend based on historical calibration records to generate a system error prediction model.

The invention relates to the technical field of visual health management, in particular to a ciliary muscle adjustment training system and method based on eye movement tracking, and the method comprises the steps: collecting eye movement data of a fixation point track, an adjustment amplitude change rate and pupil response time in real time, and combining personalized parameters such as the age of a user, diopter and a training target; inputting a ciliary muscle training dynamic planning algorithm module, dynamically generating a stage scheme containing a training action type, a single-group duration, a stimulation interval and a focal length adjustment strategy, and adaptively adjusting the strategy according to real-time adjustment stability and a historical ability curve; through multi-dimensional matching analysis of a pre-stored standard eye movement template, a training specification score is calculated from the track goodness of fit, speed uniformity, amplitude standard-reaching rate and binocular coordination, visual fatigue parameters are obtained based on continuous training duration, pupil fluctuation and retina reflex change, a scoring result is dynamically corrected, and a visual fatigue evaluation result is obtained. Personalized, self-adaptive and quantitative safety control of the training process is realized.

The invention relates to an intelligent agent continuous training and effect evaluation closed-loop method and system fusing work order feedback, and a medium, and relates to the technical field of deep learning. The agent continuous training and effect evaluation closed-loop method comprises the following steps: receiving and analyzing work order feedback information, and recording the number of error information and information storage duration in combination with a preset structure sample label; dynamically judging the number of error information and the information storage duration, triggering a retraining rule of an original agent model, distributing candidate information weights, and screening a target sample set; directionally training and optimizing the original agent model according to a retraining rule in combination with the target sample set, and generating a new agent model; constructing a historical work order problem set, comparing and verifying a new agent model in combination with an original agent model, generating an agent evaluation table, and correcting a candidate information weight; by accurately focusing the weak link, the continuous and efficient enhancement of the model capability is realized, so that the answer accuracy of the intelligent agent is remarkably improved, and the AI illusion is effectively inhibited.

The invention discloses an ultra-lightweight low-resolution dark light face enhancement method and system, and belongs to the technical field of computer vision and artificial intelligence, and the method comprises the steps: obtaining a degraded image, and processing the degraded image into a shallow feature map; inputting the shallow feature map into a hierarchical network structure; setting a module group in each level of the encoder and the decoder; only the first Transform module is used for carrying out complete self-attention calculation; the other Transform modules except the first Transform module in the module group execute attention sharing operation; inputting the deep feature map into an image reconstruction module formed by a convolutional layer to generate a residual image; and carrying out pixel-level addition on the residual image and the input degraded image to obtain a recovered clear image. According to the method, the problems of high calculation complexity and information redundancy in the prior art are solved. According to the method, the calculation overhead is remarkably reduced through an attention sharing strategy, and the excellent recovery quality is ensured through wavelet enhancement.

The invention belongs to the technical field of electric power measurement, and discloses a calibration method and system of an intelligent measurement switch and a multi-switch synchronous measurement method, and the method comprises the steps: obtaining voltage and current waveform data of the intelligent measurement switch under a nonlinear load condition, and extracting a frequency spectrum feature matrix of the voltage and current waveform data; constructing an amplitude-phase coupling error model representing a nonlinear incidence relation between amplitude fluctuation and phase deviation, inputting the spectrum characteristic matrix into the amplitude-phase coupling error model, and calculating an amplitude-phase coupling error vector; performing vector decomposition on the amplitude-phase coupling error vector, and separating out an in-phase component error and an orthogonal component error; and generating a calibration compensation coefficient according to the in-phase component error and the orthogonal component error, and performing vector calibration on the intelligent measurement switch, thereby improving the calibration precision of the intelligent measurement switch.

The application provides a consumer rights protection intelligent review method and system based on a large language model, and relates to the technical field of artificial intelligence and natural language processing. The method comprises: preprocessing a text to be reviewed to obtain structured text feature data; using a pre-trained large language model to perform semantic understanding on the structured text feature data to obtain a text semantic vector containing domain semantic information; performing joint retrieval and matching on the text semantic vector in combination with a knowledge graph and a rule base to identify and output a risk element set; and performing compliance determination based on the risk element set to obtain a preliminary review conclusion containing a risk level, evidence sufficiency, rule matching degree and benchmark confidence. The application automatically converts an original text into a high-confidence review conclusion with explainability through intelligent processing of the whole process.

The invention relates to the technical field of ship safety, and discloses a ship safety management monitoring system based on the Internet and a Beidou satellite, the system comprises a coordinate analysis module, a ship state module, a data fusion module, an abnormity identification module and an early warning message sending module, a satellite positioning signal is analyzed, and real-time latitude and longitude coordinates of a ship are obtained; generating comprehensive operation parameters of the ship according to the attitude, the navigational speed, the load and the main engine operation parameters of the ship; performing data fusion on the longitude and latitude coordinates after timestamp alignment and the comprehensive operation parameters to obtain an operation feature vector of the ship; performing abnormal mode identification on the operation feature vector based on historical navigation data and a preset safety threshold to obtain a risk category of the ship; when a risk category occurs, generating a treatment measure message of the risk category, and sending the treatment measure message to a ship-end display terminal and a shore-based management terminal through communication between the Internet and a Beidou short message; the ship safety management efficiency can be improved.

The invention discloses a wind and light storage cooperative control system based on source load matching. The system aligns a source load storage network and communication information and carries out credible evaluation; wind, light and load are predicted, and uncertainty levels are given; calculating a matching degree and a short plate hot spot, and combining network side hard constraints such as voltage, power flow and reverse transmission to obtain an acceptable boundary; when the matching is qualified and the complementary energy is still available, identifying the complementary energy / energy shortage and the reason, quickly checking the transferable margin and loss by the edge, generating a routing plan containing retention time and action budget, and scheduling the stored energy and the flexible load in a rolling manner; meanwhile, the risk budget and standby are written back according to feasibility, prediction uncertainty and communication quality, gray release and automatic rollback are achieved in cooperation with degradation and playback scoring, and consumption and traceability are improved.

The invention discloses an automatic driving test and evaluation method and device based on a large language model, and solves the problems that the pertinence of a test scene is insufficient, the result understanding depends on manpower, the evaluation interpretation is weak and the like in a traditional method. By means of semantic analysis, logical reasoning and domain knowledge support of a large language model and in combination with an automatic driving test domain knowledge graph, generation of a structured test portrait of a tested system, construction of a test demand list, generation of a semantic test scene description file, interpretable test result attribution, multi-dimensional comprehensive evaluation and report output are completed in sequence. And a test portrait is updated through report feedback to form a closed loop. The method is suitable for function verification and performance evaluation of an automatic driving simulation test, can also be used for post-processing analysis of real vehicle test data, realizes automation of a whole test process and interpretability of an evaluation result, and improves test precision and efficiency.

The invention relates to the technical field of diet health services, in particular to a diet service method and system based on human nutriology. The method comprises the steps of collecting multi-dimensional data for a user; on the basis of a human nutriology database and user full-amount data, a personalized nutrition target is determined through nutrition requirement calculation; in combination with personalized nutrition targets, user dietary habits and actual life scenes, food material recipes are matched to generate multiple sets of personalized dietary schemes; providing diet scheme execution guidance, and adjusting the diet scheme; the personalized diet scheme and the updating scheme are pushed to the user; the system comprises a user data acquisition module, a nutrition target determination module, a diet scheme generation module, a diet scheme guidance and dynamic adjustment module and a push module. According to the method, multi-dimensional adaptation of nutritional demand analysis to user data is realized, the disease state, exercise intensity and dietary habits of the user are fully combined, the nutrition target pertinence is improved, and the actual demand of the user is accurately matched.

The invention relates to the technical field of battery management, and discloses a battery health state evaluation method and system for an electric tool energy supply device, and the method comprises the steps: obtaining multi-source sensor data in the operation process of an electric tool, and carrying out the preprocessing of the data; identifying and classifying multi-stage pulse discharge events, extracting multi-scale voltage response features and performing temperature compensation; equivalent circuit model parameter identification is carried out, abnormal identification results are checked and removed, and parameter aging characteristics are extracted; constructing a time sequence prediction model for prediction; evaluating the overall health state and the monomer consistency of the battery pack; defining a judgment standard of effective cycles, predicting the number of remaining effective cycles, calculating a confidence interval of a prediction result, and generating a battery management suggestion; formatting into a standardized evaluation report, carrying out data synchronization and model optimization through edge cloud collaboration, and recording a system operation log; according to the invention, accurate health state evaluation and residual life prediction are realized.

The invention belongs to the technical field of photovoltaic power station power generation, and particularly relates to a single-crystal assembly photovoltaic power station power generation data monitoring and analyzing system, which comprises a multi-modal sensing acquisition and real-time analysis layer, a single-crystal assembly power generation data analysis layer, a single-crystal assembly power generation data analysis layer and a single-crystal assembly power generation data analysis layer, and constructing a component-level multi-physics field coupling model, wherein the model fusion method comprises the following steps: adaptively adjusting the monitoring acquisition frequency of a corresponding region, focusing key data acquisition in a fault incubation period, and keeping the conventional monitoring intensity when the component state is stable. And meanwhile, the cooperative control module for carbon flow monitoring, analysis and energizing generates a precise control instruction based on a dynamic optimization result, so that the problems of one-way open loop and collection strategy stiffness of traditional system monitoring and analysis are solved, the pertinence of fault diagnosis is improved, unnecessary monitoring resource consumption is reduced, and the system monitoring energy efficiency is improved by more than or equal to 40%.

The invention relates to the technical field of antibacterial materials, and provides a light-enzyme-heat synergistic antibacterial nano composite material and a preparation method thereof.The preparation method comprises the steps that 1, a copper-based chalcogenide with a plasma resonance effect serves as a structural template, a transition metal oxide intermediate layer and a bismuth oxyhalide epitaxial layer are sequentially constructed on the surface of the copper-based chalcogenide with the plasma resonance effect through a step-by-step continuous solvothermal method, and the copper-based chalcogenide with the plasma resonance effect serves as the structural template; forming a core-shell heterojunction with Z-type energy band matching; (2) pre-assembling a natural enzyme and a monatomic nano-enzyme in a buffer system to form a composite catalytic unit with an electron transfer channel; and (3) by taking the core-shell heterojunction as a polymerization core, under the precise temperature control condition of 0-10 DEG C, initiating active free radical polymerization through an interface, constructing a thermal response polymer shell layer on the surface of the heterojunction, and encapsulating the composite catalytic unit in a functional microdomain of a polymer network in situ. The anti-bacterial fabric can achieve a better synergistic anti-bacterial effect.

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.

This invention relates to a method for monitoring the microenvironment of the skin around the stoma and optimizing chassis adaptation, comprising: dividing the stoma chassis bonding interface into multiple circumferential detection units, and within each circumferential detection unit, dividing it radially into an inner zone, a middle zone, and an outer zone; collecting two or more of the following parameters from each zone: humidity, pH value, contact pressure, and contact impedance, and marking them as a qualified sealing zone, a weakened sealing zone, or a failed sealing zone according to preset thresholds; determining whether the weakened sealing zone and the failed sealing zone within the same circumferential detection unit extend from the inner zone through the middle zone to the outer zone and form a continuous adjacent pathway; generating local sealing compensation parameters when not formed, and generating opening gap correction parameters, sealing compensation parameters, or bonding area adjustment parameters based on the spatial position of the pathway when formed, thereby implementing differentiated processing for the formation stage and spatial position of the leakage pathway, improving the pertinence and reliability of chassis adaptation optimization.

The embodiment of the invention provides an image detection method and computing equipment. The method comprises the following steps: segmenting a first image to be detected into a plurality of image blocks; performing frequency domain analysis on each image block to obtain each frequency domain analysis result corresponding to each image block; based on each frequency domain analysis result, determining a plurality of first image blocks meeting a preset screening condition from the plurality of image blocks, the preset screening condition at least comprising: the frequency domain analysis result indicates that the richness of the contained high-frequency features is the first N, and the high-frequency features correspond to an area in which the gray scale variation amplitude exceeds a specified threshold in the image blocks; determining image features corresponding to the first image based on the plurality of first image blocks through an extraction network; and determining a detection result at least indicating whether the first image is the generated image or not through a detection network at least based on the image features so as to realize accurate detection of the generated image.

The utility model relates to concrete technical field discloses fly ash content concrete's test simulation cabin that can simulate underground environment, including cabin body, the inside bottom wall of cabin body equidistance is opened with a plurality of ponding holes, the outer wall right side of cabin body is fixedly connected with fixed plate, the top wall of fixed plate is fixedly connected with water pump, one end of water pump is communicated in the outer wall right side middle lower portion of cabin body, the other end of water pump is fixedly connected with water pipe, water pipe is communicated in the outer wall right side middle upper portion of cabin body, the inside top wall of cabin body is installed with the shower head equidistance. In the utility model, motor drive connecting shaft rotates, connecting shaft drives gear no.

The application relates to the field of sanitation operation scheduling, in particular to a sanitation operation scheduling method based on the Internet of Things, which comprises the following steps: determining an iterative stop condition of a load CEEMDAN model based on time local weighted variance of load data, so as to generate vehicle load characteristics; generating vehicle oil consumption characteristics through an oil consumption CEEMDAN model based on statistical values of load data residuals and oil consumption data residuals, so as to calculate oil consumption channel residual signals; generating comprehensive vehicle carrying characteristics through a feature extraction model based on a multilayer perception machine, a gating mechanism and a cross-modal attention mechanism; and generating corrected vehicle garbage load through a long short-term memory network and a fully connected layer architecture. The application realizes efficient noise reduction and accurate decomposition of non-stationary time series data of oil consumption and load, realizes accurate processing of Internet of Things data of garbage transfer vehicles, improves the operation efficiency of the garbage transfer vehicles and reduces the operation cost.

The application relates to the technical field of fault monitoring of wind turbines, in particular to a doubly-fed wind turbine fault monitoring method and system based on frequency monitoring, which comprises the following steps: acquiring operation parameter data of a doubly-fed wind turbine, generating a similar working condition time period of a gearbox in the doubly-fed wind turbine according to the operation parameter data based on a clustering algorithm; extracting target vibration data; generating modal components and central frequencies corresponding to each modal component according to the target vibration data, and generating modal component energy; generating a fusion modal energy value according to the modal component energy, and generating a unit fault early warning result. According to the application, only vibration data corresponding to the similar working condition time period is extracted, so that the interference of signal differences under different working conditions on vibration characteristics is excluded, a comprehensive criterion for unit faults based on the fusion modal energy value is used, the gearbox fault risk is early warned as accurately as possible, and false negatives are reduced.

The application provides a teaching intervention method, system and device based on multi-modal intelligent analysis, and relates to the technical field of intelligent education. The method comprises the following steps: collecting multi-modal learning data of a learner, extracting each modal learning feature, and inferring the current learning state of the learner; if the current learning state is abnormal, at least one main factor causing the abnormality is screened out based on the causal relationship between a plurality of preset candidate factors and the abnormal learning state, and a corresponding first set of intervenable variables is determined; the causal relationship is obtained based on learning portrait analysis including learning state history records; an intelligent agent is called, and a teaching plan and teaching intention for the current learning state are generated based on an intelligent agent data set including the learning portrait, the main factors, the first set of intervenable variables and the teaching task of the current round; and a digital person is driven to execute the teaching plan and the teaching intention. The application can comprehensively perceive the state of the learner, and realize collaborative optimization of teaching strategies and teaching interactive behaviors.

This invention relates to the field of data processing and discloses an intelligent cleaning method for multi-source heterogeneous big data. The method includes: collecting structured, semi-structured, and unstructured data from multiple source systems and parsing them to obtain raw data units; performing standard concept mapping on the raw data units and writing source identifiers, extraction method identifiers, collection time identifiers, upstream task identifiers, and derived link identifiers to generate data units with lineage identifiers; forming candidate association clusters based on stability identifiers, contextual features, and temporal adjacency relationships; constructing a conflict evidence graph within the candidate association clusters and determining edge weights by combining source credibility, time freshness, lineage propagation depth, and source drift coefficient; locating conflict-connected subgraphs and determining the minimum repair boundary by combining the data lineage graph; generating multiple candidate repair strategies and arbitrating them.

The present application relates to the technical field of environmental monitoring and pollution tracing, and discloses a basin-scale perfluorinated compound tracing method and system based on artificial intelligence, which integrates multi-source data in a basin to construct a unified spatial and temporal granularity basin multi-source data cube, identifies perfluorinated compound structures through a virtual spectrum diagram enhanced deep learning model, simulates pollutant spatial and temporal concentration distribution by combining a graph neural network of a physical mechanism, and then inversely propagates and matches source characteristic fingerprints to invert potential emission source contribution, outputs a pollution source contribution heat map and a priority control enterprise list, solves the core problems of difficulty in joint analysis of multi-source data, difficulty in identification of unknown compounds and difficulty in quantitative determination of multi-source superposition contribution in basin-scale perfluorinated compound tracing, and provides support for precise pollution control.

The invention relates to the technical field of medical robots, and particularly discloses a multi-degree-of-freedom ankle training mechanism and a simulation and movement assisting robot. The multi-degree-of-freedom ankle training mechanism comprises a base, a moving part and a telescopic driving mechanism, the moving part comprises a moving base plate, a pedal plate and a rotary driving mechanism, and the moving base plate is connected with the base through a first universal hinge; the telescopic driving mechanism is arranged on the base, and a telescopic rod of the telescopic driving mechanism is connected with the moving base plate through a second universal hinge. According to the scheme, the telescopic driving mechanism is arranged on the base, and the movement base plate and the pedal plate are driven by a simpler mechanism to rotate at multiple degrees of freedom, so that the size and movement inertia of movement parts are greatly reduced, and the overall structure is more compact and light. The motion response speed and the dynamic performance of the multi-degree-of-freedom ankle training mechanism can be improved, the multi-degree-of-freedom ankle training mechanism can be integrated into a simulation and motion assisting robot platform more easily, and the system integration degree is improved.

The application discloses a kind of optimization methods of tubular heat exchanger low-carbon manufacturing process parameters, it is related to tubular heat exchanger manufacturing technical field, including the following steps: according to process, tubular heat exchanger manufacturing process is split, is divided into multiple with carbon emission related process object, the corresponding relationship between process parameters and carbon emission is established, and the before-and-after relationship and parallel relationship between each process are described using AND / OR logic, form process association relationship.The application integrates each process carbon emission calculation by establishing multilevel mapping relationship, realizes whole-process low-carbon collaborative optimization;Combining Sobol method identifies key process parameters, improves parameter screening efficiency and rationality.At the same time, using double population and fuzzy logic regulation mechanism improves optimization ability under complex constraint, the obtained parameter can be directly used for production, realizes the comprehensive optimization of carbon emission, processing time and cost without equipment modification.

This invention relates to the field of intelligent auditing, and more particularly to a graph neural network method and system for mining hidden relationships among multiple sets of enterprise accounts. The method includes: acquiring multiple sets of accounts of a target enterprise, extracting entity and transaction information, and constructing a heterogeneous graph network containing multiple types of nodes such as enterprises, individuals, bank accounts, and transaction vouchers, as well as multiple types of edges such as fund transfers, control, and voucher associations; inputting the heterogeneous graph into a pre-trained GNN, aggregating neighbor features through graph convolutional layers, and updating node embedding vectors; based on the embedding vectors, using an attention mechanism to calculate the confidence of hidden relationships between node pairs, and filtering suspicious related node pairs; performing path reasoning and backtracking on suspicious node pairs, mining multiple paths, and generating a chain of evidence for related paths. This application, by constructing a heterogeneous graph containing multiple types of nodes and edges, comprehensively captures entities such as enterprises, individuals, accounts, and vouchers and their multidimensional relationships, solving the problem that existing methods are unable to comprehensively characterize complex relationship scenarios and capture hidden relationships.

The application relates to a vehicle underbody living body detection method, a computer program product, an electronic device and a vehicle. The detection method comprises the following steps: taking a vehicle underbody image as a reference image when the vehicle is powered off; determining a real-time image with a difference region as a to-be-detected image based on the reference image and the real-time image of the vehicle underbody; determining feature detection information of the to-be-detected image, and realizing identification of a vehicle underbody living body and / or determination of a vehicle underbody living body state based on the feature detection information. Based on this, the application can improve the pertinence and efficiency of vehicle underbody detection, and guarantee the accuracy and effectiveness of vehicle underbody living body detection results.

The invention relates to the technical field of multi-source data processing, in particular to a multi-source data coupling system and method, and the method comprises the steps: obtaining original monitoring data of a target region from a plurality of data sources, carrying out the quality evaluation of the data of each data source, and generating a quality score; establishing a data source contribution degree dynamic weight distribution model by combining geological activity stage parameters, calculating a real-time dynamic contribution weight of each data source, and performing weighted fusion on original data; introducing geological priori knowledge to perform adaptive feature screening on the fused data, retaining key features, and performing collaborative analysis on the key features and the geological priori knowledge to obtain a geological state coupling result; and constructing a closed-loop optimization mechanism, calibrating a data quality score by using a coupling result, and updating a weight model for fusion in the next period. The method can adapt to dynamic changes of geological activities, improve reliability of fused data, realize adaptive coupling optimization of multi-source monitoring data, and adapt to complex geological monitoring scenes.

This application relates to the field of industrial data processing and intelligent control technology, and discloses a method for dynamic modeling and prediction of industrial big data combined with digital twins. The method includes: acquiring operating status data, process execution data, control command records and prediction results, and constructing a sequence of control action events; segmenting the operating status data according to the sequence of control action events to form natural operating data segments and intervention operating data segments, extracting distribution characteristics, fluctuation characteristics and response characteristics, and generating a segmented behavioral feature set; calculating the control reflexive influence quantity and generating a reflexive drift marker, dividing the natural evolution dataset and the control feedback dataset, and predicting them separately; generating control risk constraint quantity, screening, delaying or limiting the control commands to be executed, and iteratively correcting the dataset according to the execution results; this method can distinguish between control feedback changes and natural evolution changes, reduce drift misjudgments, and suppress multi-round oscillations and parameter fluctuations.

The invention discloses a Tibetan language image-text emotion data multi-round consistency labeling method and a Tibetan language image-text emotion data multi-round consistency labeling system, and belongs to the technical field of natural language processing and computer vision. The method comprises the following steps: establishing a Tibetan language labeling standard knowledge base containing six types of basic emotions through a Tibetan language emotion labeling knowledge base construction module; constructing a typical language pattern feature set and a facial expression visual feature description set, and establishing an association mapping relationship between the two feature sets; multiple rounds of independent labeling are implemented through a multi-round back-to-back labeling execution module; a Fleiss Kappa consistency coefficient is calculated through a dynamic consistency monitoring module, and the consistency level is monitored; classifying the bifurcations into a semantic ambiguity type, a visual blurring type or a cultural difference type based on an association mapping relation through an intelligent bifurcation detection and expert decision module, and pushing the bifurcations to an expert for decision; and through the annotator ability portrait and a dynamic allocation module, constructing an annotator ability portrait vector and dynamically adjusting task allocation.