3349 results about "Point data" patented technology

The invention relates to the technical field of process automation, in particular to an RPA service automation method capable of enabling a large language model MCP. The method comprises the following steps: acquiring user input data, and performing intention understanding through a multi-round intention recognition model according to the user input data to obtain semantic instruction data; performing standard semantic protocol conversion according to the semantic instruction data to obtain standard semantic protocol data; performing flow chart node matching according to the standard semantic protocol data to obtain flow chart node data; and generating a business process according to the flow chart node data to obtain business process data. According to the method, the multi-round intention recognition capability of the large language model and the standard semantic protocol conversion are fused, so that the accurate mapping from the natural language to the business process is realized, and the intelligent level and the user friendliness of the RPA system are improved. Through process node matching and process logic verification, the execution effectiveness of the automatic generation process is ensured, and the labor cost and configuration errors are reduced.

The invention discloses a human body motion posture evaluation method and system. Low-cost and high-precision human body posture analysis is realized through common camera equipment. The method comprises the following steps: acquiring a human body motion video through camera equipment; detecting key points of a human body by using a deep learning model, and extracting two-dimensional or three-dimensional coordinates of the human body; calculating joint angle, speed and acceleration based on the key point data; and analyzing attitude stability, motion symmetry and fluency characteristics, and generating a comprehensive evaluation result. The system evaluates the motion quality of the user through comparison between dynamic feature extraction and a standard motion template, and generates personalized improvement suggestions based on specific indexes, including direction adjustment and amplitude range adjustment. The method can be widely applied to scenes such as exercise and fitness, rehabilitation medical treatment and competitive training, and the comprehensiveness, the accuracy and the guiding effect of action evaluation are improved. By means of the method, the user can obtain motion quality feedback in real time, the posture is effectively optimized, and the exercise performance or rehabilitation effect is improved.

The embodiment of the invention provides a kafka storage and calculation separated data processing method and device, and the method comprises the steps: constructing a multi-layer data processing architecture which comprises four layers: network communication, task distribution, processing scheduling and storage. Cold and hot data separation is realized through hierarchical design of a log pre-writing area and a main storage area, and intelligent scheduling of system resources is realized in combination with a copy manager, a coordinator and a load predictor. An adaptive compression algorithm and a verification mechanism are adopted to ensure the data transmission efficiency and integrity, a message cache region is created to store hotspot data, and flexible access of cold and hot data is realized based on message sites. According to the method, the defects of the traditional technology in storage and calculation coupling, storage efficiency, load balancing and the like are effectively overcome, and the performance and expandability of a message queue system are remarkably improved.

The invention relates to the field of artificial intelligence and computer vision, and discloses an artificial intelligence physical training exercise system which comprises a data acquisition module, a data preprocessing module, an action modeling module, an action evaluation module and a user feedback module. The data acquisition module acquires real-time video data of a moving target and extracts skeleton key point data; the data preprocessing module performs optimization processing on the skeleton key point data; the action modeling module is used for modeling and predicting the action state of the moving target by utilizing a target constraint dynamic optimization model based on the optimized skeleton data; the action evaluation module generates action scores and optimization suggestions through local action decomposition and action matching analysis; and the user feedback module provides real-time feedback information and generates a training summary report according to the action score. The method can achieve the efficient analysis and precise evaluation of the motion of the moving target in a complex scene, provides personalized optimization guidance, and is widely suitable for physical training and teaching scenes.

The invention relates to a synchronous line loss intelligent diagnosis and analysis system based on an electric power knowledge graph. The synchronous line loss intelligent diagnosis and analysis system comprises a data acquisition layer, a data access layer, a data storage layer, a data analysis layer and a visualization layer, the data acquisition layer is used for acquiring power grid node data; the data access layer is used for carrying out dynamic synchronization on power grid node data based on Apache Kafka; the data storage layer comprises a time sequence database and a graph database; the data analysis layer comprises a detection module and an inference engine; and the visualization layer is used for dynamically displaying a propagation path and node characteristics of an abnormal region based on a topological structure of the knowledge graph, and querying an abnormal reason of the node and the influence on the whole system by clicking the node of the device. According to the invention, intelligent detection and diagnosis of the line loss abnormal area are realized, and the problem root cause is rapidly identified.

The invention provides a wharf facility health monitoring method and system based on digital twinning and machine learning, and relates to the field of wharf facility health monitoring, and the method comprises the steps: building a weight coefficient of a monitoring point network topological graph according to a synchronism change value of the same type of sensor data corresponding to adjacent monitoring points; determining an abnormal transmission path based on the weight coefficient; determining an abnormal state propagation rule of each abnormal transmission path in each preset abnormal state according to the historical monitoring data corresponding to each preset abnormal state; after abnormal sensor data of any monitoring point is detected, parameter predicted values of other monitoring points in the corresponding abnormal transmission path are calculated based on an abnormal state propagation rule; and if the parameter prediction values of the preset proportion number deviate from the corresponding actual monitoring values, generating early warning information. By implementing the method, false alarms caused by single-point data fluctuation summarizing in a traditional health monitoring mode can be dealt with, and the accuracy and reliability of abnormal state early warning are improved.

The invention relates to the technical field of image recognition, in particular to an image processing method and system for rehabilitation training action analysis. According to the method, a multi-view image sequence is collected based on a binocular camera device, skeleton key point data of a user in a training process is extracted by utilizing a three-dimensional attitude reconstruction technology, and an action three-dimensional time sequence data set is constructed. The method comprises the following steps: firstly, constructing an individual standard power generation characteristic model of a user, modeling a skeleton driving path of a main muscle group, and forming a personalized power generation reference structure; and then a standard rehabilitation action path is matched through a dynamic time warping algorithm, and the attitude deviation under the key frame is identified. And the system dynamically compares the identified non-standard motion mode with the individual model, judges whether abnormal force generation exists or not and outputs the type and the part of the muscle compensation behavior. And finally, multi-modal feedback information with highlighted graphs, voice prompts and character suggestions is generated in combination with an identification result, so that the identification precision and personalized guidance capability of rehabilitation training are remarkably improved.

The invention discloses a self-adaptive three-dimensional scene reconstruction method and system based on a single panorama, and belongs to the field of computer vision processing. The method comprises the following steps: firstly, generating a depth map through an indoor scene panorama and constructing an initial three-dimensional grid; then generating a multi-view image and a shielding mask thereof based on view conversion, forming a training sample pair for fine tuning of the diffusion completion model, and injecting scene prior information for the model; then extracting a grid boundary contour and calculating a central axis, and adaptively constructing a camera pose set; after the integrity of the grid is optimized through iteration completion, the grid is converted into a 3D Gaussian sputtering field, and Gaussian point parameters are optimized through multi-view data; in the optimization process, an up-sampling strategy of rendering error feedback and edge Gaussian points is adopted, and finally a scene model with complete geometric textures is output. According to the method, a high-quality three-dimensional scene is reconstructed from a single panorama through a structure self-adaptive completion and Gaussian refining technology, and the method is particularly suitable for immersive roaming reconstruction of indoor scenes.

The invention discloses a building construction process risk assessment method and system based on a fault tolerance mechanism, and particularly relates to the field of building construction process risk assessment, and the method comprises the steps: building a measurement point hierarchical index of a building construction process, synchronously arranging time sequence displacement data, and executing difference calculation to obtain a space gradient evolution sequence; mapping a trend change interval set extracted by sliding window processing into a path node set, and constructing a topological connection relation of the path node set to form an initial risk path map; the data state of each path node in the risk path map is subjected to anomaly identification, the access control effect of the abnormal node in the path is analyzed, and whether the path fracture risk is formed or not is judged in combination with the sensitive monitoring area. By constructing a risk map structure based on a trend path and combining a path interruption node identification and alternative path dynamic deduction mechanism, continuous assessment and map self-correction of the risk trend under the condition of measuring point data failure are realized.

The invention discloses a data enhancement method and system based on multi-agent self-evolution and hybrid evaluation, and relates to the technical field of artificial intelligence, and the method comprises the steps: employing a teacher model as a multi-agent cooperation system, and carrying out the interaction of a reasoning agent, an evaluation agent, a reflection agent and a memory management agent through a reasoning agent, an evaluation agent, a reflection agent and a memory management agent; iteratively generating a high-quality reasoning data set with a traceable reasoning path and a self-verification label, and performing multi-task supervision fine tuning on a learning model; a diversity sampling strategy based on reinforcement learning is adopted to generate multiple groups of outputs, and a weak point data set is screened by using a consistency score; and for the weak point data set, the multi-view instruction rewriting agent performs diversity rewriting on the instruction and then returns to perform deep reasoning distillation to generate new enhanced data and combine the new enhanced data to high-quality reasoning data. According to the method, multiple agents are arranged, so that the model is iteratively synthesized, evaluated, reflected and modified, the obtained enhanced data can be directly used, a manual auditing step is omitted, and the synthesis efficiency is improved.

The invention discloses an intelligent water affair monitoring and management system based on digital twinning, and relates to the technical field of intelligent water affair. The intelligent water affair monitoring and management system comprises a water affair monitoring and management platform, and the water affair monitoring and management platform is in communication connection with the following modules: a multi-source data acquisition module; the water affair monitoring system is used for collecting and preprocessing water affair monitoring data from a plurality of monitoring points of the water affair system, monitoring changes of a pipe network topological structure and obtaining dynamic data of a pipe network connection relation and geometric parameters. Through the digital twinborn technology, data of multiple monitoring points can be integrated in real time, abnormal events such as water quality pollution, equipment faults, water shortage and hydraulic change can be rapidly recognized in combination with the abnormal trend analysis module, early warning signals are automatically generated, the response time is remarkably shortened through an instant early warning mechanism, and the early warning efficiency is improved. Therefore, the management personnel can take measures at the initial stage of the abnormal event, the problem expansion is effectively prevented, and the timeliness and accuracy of water management are improved.

The invention discloses a sensor fault detection and reconstruction method based on multi-sensor fusion, and the method comprises the steps: collecting data in a production process, and obtaining the data information of a detection point; carrying out anomaly detection on the collected data information through an LOF algorithm; classifying the detected abnormal data into normal data and abnormal data; inputting the abnormal data into a CNN-LSTM network sensor fault diagnosis model for principal component analysis dimensionality reduction, and reducing the input abnormal data from high-dimensional data to low-dimensional data by using principal component analysis; inputting sensor abnormal data subjected to principal component analysis dimensionality reduction into a CNN-LSTM-based sensor fault classification model, and determining the type of a sensor fault; and reconstructing the data of the fault sensor.

The invention relates to a power distribution network distributed energy bearing capacity assessment method and device. The method comprises the following steps: carrying out topology analysis on a network structure of a power distribution network to obtain an initial network topology model; obtaining a node dynamic feature data set based on the initial network topology model and the distributed energy access point data of the power distribution network; wherein the node dynamic characteristic data set comprises operation parameters of each node of the power distribution network in different load scenes; generating a parameter incidence matrix according to the node dynamic characteristic data set, and obtaining a bearing capacity reference model of the power distribution network according to the parameter incidence matrix and real-time data of the power distribution network in an operation state; wherein the parameter incidence matrix is used for quantifying the coupling degree between the operation parameters; and obtaining a risk distribution mapping graph according to the bearing capacity reference model, and identifying a potential overload area of the power distribution network based on the risk distribution mapping graph. According to the invention, power distribution network operation risk assessment can be accurately realized.

The invention relates to the technical field of subtitle translation, in particular to a multi-modal time sequence alignment AI video translation method and system, and the method comprises the steps: 1, carrying out the multi-modal analysis of a to-be-translated video, and obtaining audio separation data, voiceprint feature data and visual time sequence data; 2, performing cross-language translation and context optimization on the basis of the voice of the audio separation data to generate a target language text, and synthesizing target language voice retaining the original voice color in combination with the voiceprint feature data and the target language text; generating a mouth shape animation matched with the target language voice based on the lip key point data and the limb action time sequence data; and step 3, performing four-dimensional alignment on the target language voice, the translated text, the mouth shape animation and the limb action sequence through a cross-modal time sequence encoder, and dynamically adjusting the layout of the bilingual subtitles to adapt to a video picture. According to the method and the device, multi-mode synchronization can be taken into consideration during video translation, so that the body actions such as voice, subtitles and mouth shapes are kept aligned.

The invention provides a human body skeleton point positioning identification method and system based on OpenPose, and the method comprises the steps: carrying out the preprocessing of an input human body image, and obtaining the preprocessing image data; performing skeleton point detection on the preprocessed image data by adopting a double-domain multi-path self-supervised diffusion model in combination with a convolutional neural network to obtain two-dimensional human skeleton point coordinate data; performing signal-to-noise ratio evaluation and noise reduction processing on the two-dimensional skeleton point coordinate data to obtain two-dimensional skeleton point data after noise reduction; performing two-dimensional to three-dimensional conversion through a triangulation principle and a feature matching technology to obtain preliminary three-dimensional skeleton point coordinate data; performing skeleton point optimization through rate perception analysis and a three-dimensional Gaussian compression algorithm; and carrying out shielding prediction and completion processing on the optimized three-dimensional skeleton points to obtain a complete human skeleton point positioning identification result. According to the invention, the problems of accuracy and stability of skeleton point positioning in a complex dynamic scene are solved.

Provided is a method for perceiving and eliminating an abnormal state of active distribution network based on data enhancement, including: acquiring, by synchrophasor measurement device, data of each node of active distribution network in target domain in real-time and transmitting to processor; inputting the acquired data into a classification model, and outputting abnormal detection and classification results in real time; and analyzing the abnormal detection and classification results, and transmitting an abnormal state eliminating instruction to a distribution terminal to eliminate the abnormal state. Wherein, hidden distribution features in node data of active distribution network are mined through dynamic clustering, a large amount of unlabeled data are clustered, a data label is generated through self-coding and label correction rule, training samples with balanced category distribution is generated through data enhancement and is used to train the classification model based on dynamic graph attention network by domain adaption method.

The invention provides a 3D scene reconstruction method and system based on adaptive key frames, and relates to the technical field of computer vision, and the method comprises the steps: obtaining continuous multi-frame images and inertial measurement unit data of a current scene; acquiring camera trajectory data and offset data of each inertial measurement assembly based on the inertial measurement unit data; carrying out target feature extraction and target tracking on each frame in the multiple frames of images to obtain feature point data of the to-be-tracked target; determining a key frame based on a scene complexity index and a system resource utilization rate index of the current scene; performing visual inertia joint optimization based on all feature point data, camera trajectory data and offset data, and generating point cloud data containing each feature point; and generating 3D Gaussian distribution of the current scene based on the point cloud data and the key frame, and rendering the distribution to obtain a first 3D rendered image of the current scene. According to the method, real-time 3D scene reconstruction can be carried out in a complex and changeable application scene.

The invention discloses a human body posture reconstruction and tracking method based on a neural network, and relates to the technical field of posture reconstruction and tracking. Target echo data are acquired by adopting a multiple-input multiple-output (MIMO) antenna array, and a low-resolution three-dimensional radar image is constructed by adopting space-time weighting in combination with wavelet transform noise reduction and self-adaptive gain compensation; in the feature extraction stage, a Transform-based feature extraction network is introduced, short-distance joint details are processed in combination with a local convolutional neural network (CNN), the spatial correlation of a far-end joint is calculated by adopting a global self-attention mechanism, and a global feature enhancement (GFE) module performs feature completion on a low-confidence region; a hybrid memory network HMN is adopted to carry out short-time and long-time information fusion, a short-time memory unit STM is used to store current frame features, a long-time memory unit LTM is used to store historical key point data, a drift error is corrected in combination with time sequence alignment, and time sequence consistency of attitude estimation is guaranteed.

Real-time monitoring and alerting of breached security credentials for server-based endpoints is described herein. In various embodiments, a software component (that intercepts outgoing network traffic from a computing device) or a server-based breach monitoring component may receive a request for endpoint data. Credential data may be included in the request. The endpoint and / or the credential data may be compared to credential data linked to a list of compromised endpoints. An alert may be generated for the client device when both a match is detected with credential data of a breach object on the list of compromised endpoints and the breach time field of the breach object is after a previous credentials change for the endpoint. Access to the endpoint may be blocked and / or a change password mechanism may also be displayed to change the user credentials prior to revisiting the endpoint.

The invention relates to a quick alignment method for optical coupling equipment, which comprises the following steps of: acquiring real-time position information of a laser transmitting end and a laser receiving end, acquiring relative position data by adopting a fiber grating technology, setting a dynamic compensation algorithm according to a light spot offset and a position deviation value, generating an adjustment parameter through the dynamic compensation algorithm, and aligning the adjustment parameter to the optical coupling equipment. According to the method, the structural design of optical coupling equipment is optimized, a high-precision light spot imaging module and a fiber grating positioning module are integrated, and a multi-point data fusion technology and an optical path alignment model in a dynamic environment are utilized, so that the measurement precision is further improved, and the influence of environmental interference is reduced. According to the invention, the problems of real-time performance and accuracy of light path alignment in kilometer-level laser measurement are effectively solved, and a new solution is provided for the development of a long-distance laser measurement technology.

The invention discloses an electric power system simulation scene modeling method based on big data, which comprises the following steps: collecting preset scene data and monitoring data of an electric power system, and preprocessing the preset scene data and related data; performing power distribution network aggregation on the monitoring data to obtain equivalent load node data, and performing high-dimensional classification on the equivalent load node data according to the preset scene data to obtain node classification; the node classification comprises a normal scene and an extreme scene; performing time sequence causal feature extraction on the equivalent load node data according to the node classification to obtain time sequence causal features, and constructing a power system simulation scene model based on a Bayesian network according to the time sequence causal features; and performing uncertainty evaluation on the power system simulation scene model according to a simulation error to obtain uncertainty, performing parameter layered online correction on the power system simulation scene model based on a physical model according to the uncertainty, and outputting a target model.

The invention relates to the technical field of intelligent driving, and discloses a user service recommendation method for a vehicle-mounted scene, and the method comprises the steps: collecting the vehicle-using behavior data of a user, carrying out the correlation marking of the intention of the user through a multi-dimensional marking system, carrying out the vectorization, and storing the vectorized intention into a vector database; capturing real-time behavior data of a user to generate a real-time behavior vector; based on the vector database, generating an intention context corresponding to a real-time behavior vector through a hybrid retrieval strategy; and generating a prompt project based on the intention context, inputting a large model output service recommendation result, and pushing the result to the user. By integrating mechanisms such as data annotation, vector retrieval and large model reasoning, the problem of long cold start period depending on long-term buried point data learning in the prior art is solved, potential intentions are quickly matched, the system effective time is shortened, and the service recommendation accuracy based on the user intentions is effectively improved. The invention further discloses a user service recommendation device for the vehicle-mounted scene, the vehicle and a computer readable storage medium.

The invention discloses an intelligent method and system for multi-source heterogeneous data fusion management, and belongs to the field of data processing, and the intelligent method for multi-source heterogeneous data fusion management comprises the following steps: deploying a sensor network in a historical ponding area based on historical ponding point data and urban hydrological geographic information, acquiring real-time monitoring data, wherein the real-time monitoring data comprises rainfall intensity, road waterlogging depth, community waterlogging condition and drainage pipe network water level / flow; compared with the prior art, the method has the beneficial effects that sensors are arranged in the historical ponding area to obtain real-time monitoring data, and weather forecast, a terrain model and a pipe network model are combined to construct a waterlogging prediction model to predict whether a ponding risk exists in the historical ponding area in future set time (such as 0.5-3 hours in the future); the ponding risk is treated in advance, so that the problems of traffic interruption caused by ponding, citizen travel obstruction and the like are solved.

The invention relates to the technical field of storage, in particular to an SSD wear leveling method and device based on dynamic data migration and a medium. The method comprises the following steps: monitoring access frequency, recording read-write times of a logic block in real time, and identifying hotspot data; tracking a wear state, counting the erasure times of the physical flash memory block, and evaluating the wear degree of the physical flash memory block; data migration decision: determining whether to migrate data based on the access frequency and the wear frequency; preferentially migrating data with high access frequency and high wear degree according to access frequency and wear program sorting; and data migration execution: migrating the hotspot data from the high-wear area to the low-wear area, and updating the mapping relation. According to the method, the data storage position can be intelligently adjusted according to the real-time data access characteristics and the abrasion state, and therefore the service life of the SSD is more efficiently prolonged.

The invention is suitable for the technical field of artificial intelligence, and provides a TOF camera hardware-based personnel falling intelligent monitoring method and system, and the method comprises the following steps: collecting three-dimensional point cloud data of a monitoring region through a ToF area-array camera, carrying out the data preprocessing and background subtraction of the three-dimensional point cloud data, and obtaining the point cloud data of a human body; personnel existence detection and personnel static detection are carried out based on the three-dimensional point cloud data; performing posture recognition, performing normalization conversion on the human body point cloud data to obtain a human body depth map, and performing posture detection on key points of the human body depth map to obtain human body key point data; and predicting the human body key point data based on the trained LSTM model to obtain a person posture state, and when the person posture state is that the person falls down, generating early warning information. The method is wide in application range and suitable for personnel falling detection in public areas and private spaces, the point cloud data only comprises the attitude key points of the personnel, and personnel falling reasons can be known under the condition that privacy is not invaded.

The invention relates to the technical field of automatic operation and maintenance, in particular to a visual operation and maintenance data analysis system which comprises a node data collection module, an index sorting and screening module, a multi-parameter linkage recognition module, a resource node screening module and an event track tracing module. And analyzing operation data of each node, and grouping node types according to time labels. According to the method, the operation data of different nodes are subjected to hierarchical grouping and archiving, dynamic linkage merging of various operation and maintenance parameters is realized in combination with time labels, operation and maintenance concerned objects are automatically screened under a multi-index change trend, and accurate screening of abnormal nodes and resource redundancy nodes is completed according to a comprehensive feature track. The operation and maintenance data is promoted to be converted from single-point monitoring to full-link linkage analysis, and the intuition and the globality of operation and maintenance event response are improved in the visual archiving and tracing process, so that systematic closed-loop monitoring of node states, resource allocation and operation and maintenance risks is promoted.

The invention discloses an intelligent data storage method based on a cloud platform, relates to the technical field of data storage management, and is used for solving the problem that cold and hot data identification is lagged. According to the method, the access behavior graph is constructed, the concurrent access coupling weight is calculated, the data object with high access frequency and deep path hierarchy is identified, the structure adjustment instruction is dynamically generated, the access link complexity and index hierarchy redundancy are reduced, the response delay and the popularity prediction error before and after the structure adjustment are combined, and the access efficiency is improved. According to the method, the scheduling efficiency of the hotspot data is improved, the adaptability to the change of a storage access mode is enhanced, the dynamic adjustment capability and the structure response performance are good, and the method can be used for realizing the closed-loop linkage of the access behavior perception, the structure reconstruction and the model optimization. Therefore, the method is more suitable for a large-scale data intelligent management scene under a cloud platform.

The invention provides a full-cycle intelligent education evaluation system based on a knowledge graph, and relates to the technical field of data processing, and the system is used for carrying out the correlation analysis of a learning behavior data set and a knowledge point data set of a student, calculating the mastering degree of the student on each knowledge point, recognizing the relation between the student and each knowledge point, and carrying out the calculation of the mastering degree. The method comprises the following steps: acquiring knowledge points of students, mapping the knowledge points into a graph structure to obtain a staged knowledge graph, updating the staged knowledge graph, adjusting the relationship between the students and each knowledge point to obtain a dynamic knowledge graph, extracting the real-time mastery degree of each knowledge point according to the dynamic knowledge graph, and periodically analyzing the knowledge points in combination with time dimensions. And obtaining periodic evaluation data, analyzing the mastering degree, the progress trend and the weak knowledge points of the students at each knowledge point, and generating an individual evaluation data set. According to the invention, the knowledge graph can be constructed to periodically evaluate the learning condition of the student.

The invention discloses a gesture recognition method and device based on deep learning, and the method comprises the steps: collecting a video stream, and obtaining a hand key point data set in the video stream; performing data enhancement and preprocessing on the hand key point data set; establishing a gesture recognition model based on deep learning, and performing training optimization; and carrying out lightweight processing on the trained gesture recognition model, and outputting a gesture recognition result in real time based on the lightweight gesture recognition model. According to the method, through pre-segmentation of the video image, a progressive polynomial attenuation pruning strategy, collaborative optimization of pruning and quantification and a comprehensive callback mechanism, real-time, accurate and efficient operation of a gesture recognition model on mobile equipment and an embedded system is achieved, and the high-standard requirement of a modern intelligent interaction system is met.

The invention relates to the technical field of Internet, and discloses an Internet hotspot data mining system and method based on artificial intelligence, and the system comprises a data collection module, a semantic understanding module, a dynamic clustering module, a popularity evaluation module, a visual output module and a feedback optimization module. By setting a multi-source calibration unit, when cross-platform hotspot data acquisition is carried out, standardized processing of multi-source heterogeneous data acquisition is ensured by establishing a dynamic semantic feature library and configuring adaptive weight parameters for different data platforms; meanwhile, by monitoring the semantic offset of the collected data in real time, characterization deviation generated during cross-platform data collection can be detected and eliminated, the accuracy of hotspot clustering similarity calculation is guaranteed, hotspot recognition errors are reduced, and by deploying a semantic tracking engine, when hotspot event propagation analysis is carried out, the analysis accuracy is improved. The deviation degree of event core elements is calculated by constructing spatio-temporal feature vectors, and whether topic semantics are migrated or not is judged in real time.