829 results about "Trend analysis" patented technology

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 an urban water supply management data trend analysis method based on space-time analysis, and relates to the field of data processing, and the method comprises the steps: collecting data in real time through an urban water supply pipe network sensor network, building a space-time unified coordinate system, building a space-time Kriging interpolation model based on pipe network topology, and achieving the space-time alignment of multi-source data; dividing an adaptive space-time grid by using a Voronoi diagram and a sliding window mechanism, and calculating multi-dimensional features; constructing a dynamic space-time diagram by taking a grid as a node, performing multi-step prediction in combination with a space-time diagram convolution circulation network, fusing a Kriging interpolation result, and evaluating an abnormal probability and a confidence interval through a Bayesian neural network; a monitoring layer, a prediction layer and a risk layer are overlaid in a three-dimensional GIS, a dynamic thermodynamic diagram is generated, an early warning path is optimized based on a Dijkstra algorithm, and a minimum risk topology path is output. The method has the advantages that through space-time analysis and accurate prediction, the intelligence, stability and emergency response efficiency of urban water supply management are remarkably improved, and powerful support is provided for smart city construction.

The invention relates to the technical field of intelligent control, and discloses a self-adaptive temperature curve control system for high-temperature alloy vacuum casting, which comprises a sensing module, a control module and a control module, and is characterized in that the sensing module acquires multi-point temperature information and preprocesses the information to obtain temperature distribution in a furnace; the acquisition module obtains vacuum environment parameters. The simulation module establishes a three-dimensional dynamic coupling model based on heat conduction, radiation, convection and phase change heat effect, and obtains a real-time temperature evolution prediction result. The generation module obtains a target temperature curve of the corresponding stage. The prediction control module collects a real-time temperature evolution prediction result and a target temperature curve, deviation calculation and trend analysis are carried out, and a heating power and cooling flow control instruction is generated. And collecting product feedback errors and carrying out self-learning correction on prediction model parameters to obtain an optimized control output result. The control and optimization of the high-temperature alloy vacuum casting temperature field are realized, and the structure uniformity of castings and the quality stability of finished products are improved.

The invention provides a patient vital sign anomaly detection method based on an artificial intelligence technology, and relates to the technical field of data processing, and the method comprises the steps: collecting original sign data of a patient; calculating multi-dimensional characteristic parameters; establishing an individual baseline model, and determining a comprehensive reference interval in the model; the vital sign features monitored in real time are constructed into multi-dimensional feature vectors, the multi-dimensional feature vectors are input into the individual baseline model, and the deviation degree of the real-time multi-dimensional feature vectors in the comprehensive reference interval is calculated; performing a clustering analysis to identify an anomalous aggregation region; performing trend analysis, calculating change direction consistency and continuous change amplitude of the multi-dimensional feature vector, and generating a trend analysis result; calculating an accumulated change index in a continuous time window according to a trend analysis result to obtain a dynamic confidence score; when the dynamic confidence score continuously exceeds an adaptive threshold value, determining that a vital sign abnormal event exists; the autonomy and accuracy of the method for detecting the vital sign abnormity of the patient are improved.

The invention discloses an elevator generation type health state assessment method and system based on knowledge enhancement retrieval, and the method comprises the following steps: 1) collecting multi-source operation and maintenance data of an elevator, constructing an elevator health state knowledge graph containing equipment, parts, faults, health indexes and maintenance standards, and forming a structured health assessment knowledge base; 2) performing semantic analysis on operation indexes and natural language task description based on a large language model, and generating a health evidence sub-graph in combination with knowledge graph retrieval; 3) carrying out quantitative calculation on the component risk value and the complete machine comprehensive risk value under a rule engine and multi-agent framework, and constructing a memory graph to carry out historical trend analysis and similar case retrieval; and 4) generating an interpretable elevator health state evaluation report in combination with the logical reasoning graph and the standard knowledge graph. According to the method, the accuracy and traceability of elevator health assessment can be improved, and technical support is provided for predictive maintenance of the elevator.

The embodiment of the invention provides an energy storage system fault processing method and device, equipment, a medium and a product, and relates to the field of photovoltaic power generation, the method solves the problem of low fault processing efficiency in the prior art by constructing a cloud-side-end cooperative fault processing architecture, changes the traditional static threshold judgment logic, and improves the fault processing efficiency. A decision-making mode combining multi-source information fusion and dynamic prediction is adopted, firstly, preliminary diagnosis of real-time data is achieved through edge calculation, the instantaneity of response is ensured, and then historical trend analysis, equipment health prediction and multi-dimensional real-time state evaluation are fused through a cloud platform. According to the method, the dynamic fault evaluation capable of accurately reflecting the actual severity and development trend of the fault is generated, so that the system can be adaptive to equipment aging and environment change, and a grading response strategy accurately matched with the fault grade is triggered, thereby realizing the crossing from passive alarm to active and accurate fault management and control, and improving the fault processing efficiency.

The invention discloses a switch cabinet partial discharge on-line monitoring method based on multi-parameter fusion identification, and relates to the technical field of switch cabinet partial discharge monitoring, and the method comprises the steps: collecting local signals of each switch cabinet, and extracting a multi-parameter feature vector set; establishing a dynamic environment reference model based on the feature vector set, outputting a background reference line, and generating an environment compensation amount and a self-adaptive alarm threshold value; performing compensation calibration on each cabinet feature vector set by using the environmental compensation amount, obtaining a transverse difference degree through transverse multi-cabinet comparison, and identifying potential abnormal cabinets; performing longitudinal trend analysis on the potential abnormal cabinet body to obtain a longitudinal trend result; fusing the transverse difference degree and the longitudinal trend result to obtain a comprehensive risk index, and performing state judgment and dynamic adjustment in combination with a self-adaptive alarm threshold value; according to the invention, the problems of partial discharge false alarm, missing alarm and inaccurate identification of potential abnormal cabinet bodies are effectively solved.

Hierarchical modelling and advanced feature engineering discover abnormalities in time series data with irregular trends. Data is collected in real time to ensure temporal integrity in the invention. Extraction filters and isolates useful data. Data cleansing removes noise and extraneous data after preliminary analysis identifies patterns and abnormalities. Feature engineering organizes cleansed data for machine learning algorithms. Primary storage stores this data for fast retrieval and extensive trend analysis. Holidays and weekends provide unique patterns in trend analysis. These trends are used to cluster data and create hierarchical predictive models, starting with a first-order model for general trends and increasing in order to refine residuals. Serializing these models improves storage and retrieval. Trend clusters are created from new data points, and algorithms detect pattern deviations. Statistical tests and machine learning classifiers identify anomalies and create alerts and remedial measures. The system monitors and analyzes incoming data to detect anomalies.

The invention discloses a new energy automobile charging pile multi-dimensional data fusion monitoring and early warning method and system, and relates to the technical field of new energy automobile charging safety monitoring and control. The new energy vehicle charging pile multi-dimensional data fusion monitoring and early warning method comprises the following steps: S1, constructing a charging state original data set, and preprocessing the charging state original data set; s2, performing early warning grade division in combination with a disturbance trend analysis result and discriminator output; s3, a response control condition is matched according to the early warning level state; and S4, recording the operation structured data of the charging pile, constructing an event chain, and comparing the current and historical features to identify the coincidence degree. The problem that due to the complex operation working conditions of high-frequency plugging, large day and night temperature difference and heavy environmental dust of a new energy automobile charging pile in an urban fast charging station, an existing monitoring method cannot achieve multi-dimensional sensing and dynamic adjustment, and then abnormal early warning lags is caused is solved.

The invention relates to the technical field of forest fire management, in particular to an integrated forest fire intelligent analysis method and system. The system comprises a background service system, a central intelligent analysis system and a mobile intelligent analysis system. The background service system is deployed in a cloud computing server cluster and is used for uniformly storing personnel information, fire scene information and geographic information data and realizing multi-source data fusion and authority management through a standardized API (Application Program Interface); the central intelligent analysis system is deployed in a high-performance computer, integrates a forest fire danger auxiliary decision-making model, a forest fire spreading trend prediction model, a force distribution dynamic display module and a collaborative plotting module, and supports fire spreading prediction and real-time decision-making instruction issuing based on multi-dimensional data such as weather, terrain and vegetation; the mobile intelligent analysis system realizes fire spreading trend analysis, mobile plotting and instruction receiving in an offline environment, supports offline data caching and calling in an emergency scene, and effectively improves forest fire emergency response efficiency and resource scheduling accuracy.

The invention discloses an industrial data analysis and mining system based on multi-source fusion, which relates to the technical field of information processing, and comprises a time-frequency baseline construction module, an energy coupling modeling module, an anomaly identification and verification module, a pseudo peak shielding and filtering module, a trend node extraction module and a closed-loop regulation and control execution module, a unified time base line and a cross-source phase reference field are obtained, a millisecond-level phase comb index is constructed, a zero phase anchor point is locked, and reversible time-frequency coordinates are generated. According to the method, a unified time-frequency coordinate system is constructed, accurate alignment of multi-source data is realized, a false trend is identified in combination with energy coupling modeling and anti-fact verification, false features are screened out through a rejection domain, reverse energy intervention is implemented based on credible trend nodes, an identification-verification-regulation closed loop is formed, and trend analysis robustness and regulation reversibility are improved.

The invention relates to the technical field of network communication, in particular to a 5G-A network slice dynamic adjustment and resource allocation system oriented to low-altitude economy. The system comprises a mobility management feature integration module, a network slice service level protocol evaluation module, a mobility management demand prediction benchmark construction module, a multi-dimensional trend prediction module, a joint optimization decision generation module and a strategy verification and execution module, and is used for acquiring a mobility management feature data set of a wireless network; performing network slice service level protocol evaluation on the wireless network to obtain a network slice state evaluation result; constructing a mobility management demand prediction benchmark; performing multi-dimensional trend analysis on the network slice state evaluation result to obtain a prediction result; performing joint optimization decision on the standard network slice configuration strategy to obtain a network slice adjustment strategy; and executing the verified slice adjustment strategy to realize dynamic adjustment and resource allocation of the network slice. According to the invention, the efficiency of network slice dynamic adjustment and resource allocation can be improved.

The invention discloses a method and a system for evaluating movement of two lower limbs of a stroke patient, and relates to the technical field of movement function evaluation. The invention discloses a stroke patient double-lower-limb movement evaluation system which comprises a task guide module, a movement acquisition module, a movement scoring module, an interference evaluation module, a trend analysis module, a combination analysis module and a result output module. According to the method, the scoring model based on crowd category feature matching is constructed, exercise ability scoring is performed by adopting a machine learning algorithm or a fixed weighted combination, personalized evaluation is realized according to the group features of different patients, and the generalization ability and prediction accuracy of the scoring model are improved; group modeling is carried out by introducing features such as patient age, gender and motion level, and a scoring strategy is automatically switched according to the number of samples, so that the scoring mechanism also has reliability and clinical consistency in a family scene with limited data volume.

The invention discloses an online order full-process automatic processing system and method. The system comprises an order state analysis module, a cross-platform collaborative adaptation module, a resource scheduling planning module, an environment load evaluation module and a task scheduling optimization module. According to the invention, by analyzing the trend of the key time sequence change of the order life cycle, the circulation bottleneck can be identified and the time section exceeding the threshold value can be generated, so that the task sequence is matched with the system load, the scheduling coordination and the response precision are enhanced, the execution capability is dynamically compared to form the optimal distribution instruction, and the cooperation efficiency and the transmission continuity are improved. Based on a utilization rate change identification trend, resource distribution is balanced, scheduling flexibility and controllability are enhanced, resource and energy consumption cooperative adjustment is realized through environment load evaluation, operation stability and environment adaptability are improved, high efficiency and continuous optimization of an order flow are realized, and processing capability and operation toughness of a system under high concurrency are enhanced.

The invention relates to the technical field of debugging systems, in particular to an adaptive debugging system and method for touch screen FW parameters, and the system comprises a signal acquisition module, a trend triggering module, a temperature compensation calculation module, a window generation module and a parameter output module. According to the invention, the edge of the driving voltage is carefully collected, the sliding difference algorithm is introduced to filter out high-frequency noise, meanwhile, real-time sampling of environment temperature and baseline capacitance adjustment are combined, the influence of external variables on signals can be dynamically corrected, and accurate judgment of capacitance fluctuation is improved by utilizing periodic delay deviation and regression trend analysis. By means of main frequency feature extraction and zero crossing point sudden change detection, threshold setting can be refined, self-adaptive window expansion can be achieved, touch signal response is kept sensitive and consistent all the time, adjustment and optimization can still be automatically completed even under the condition that the temperature changes frequently or signal interference is complex, parameter drift and debugging failure are effectively prevented, and user experience is improved. And the parameter automatic matching efficiency and the finished product consistency in batch production are improved.

The invention relates to the technical field of piano teaching and computers, discloses a multi-dimensional scoring auxiliary system and method in piano teaching, and aims to solve the problems that existing piano teaching is single in evaluation dimension, one-sided in data acquisition and lagged in feedback. According to the method, a five-dimensional scoring model covering rhythm, strength, timbre, posture and fingering is constructed by synchronously collecting a key triggering time sequence, an audio frequency spectrum, limb posture and finger joint tracks, and quantitative scores are output through time alignment, feature extraction and neural network fusion calculation. The system comprises a key sensing module, an audio acquisition module, a posture capture module, a score calculation module and an AR visual feedback module, and supports teaching suggestion generation, historical trend analysis, difficulty self-adaption and multi-person comparison functions. Through multi-modal data closed-loop evaluation and intelligent intervention, the teaching accuracy and individuation level are remarkably improved, and scientific quantification and efficient improvement of the piano playing ability are achieved.

The invention discloses a crested ibis habitat ecological quality trend analysis method and system, and relates to ecological quality monitoring. The method comprises the following steps: S1, constructing an intelligent sensing network, synchronously obtaining multi-source data of a habitat, identifying activity events of crested ibis, and generating a multi-dimensional habitat parameter table; s2, collecting environmental samples, and generating a microbial functional gene abundance matrix through metagenome sequencing and bioinformatics analysis; s3, taking the activity events of the crested ibis as behavior tags, and generating habitat function health indexes by coupling the parameter table and the matrix training machine learning prediction model; s4, performing spatial interpolation and trend analysis based on the habitat function health index to generate an ecological quality space-time evolution graph; and S5, based on the ecological quality space-time evolution graph, performing quantitative analysis by using a spatial differentiation statistical model, and generating a trend analysis report. By fusing multi-source data, real-time dynamic evaluation of habitat ecological quality and quantitative analysis of driving factors are realized, and a direct decision basis is provided for accurate protection.

The invention relates to the technical field of optical network operation and maintenance, and particularly provides an optical network fault prediction method and system based on MEMS characteristic embedding and digital twinning, and the method comprises the steps: collecting multi-physics field coupling parameters of optical network operation in real time through an MEMS multi-physics field coupling sensing unit which is disposed at an optical network key node and is embedded in a physical layer device; constructing an optical network cross-scale digital twin, inputting parameters to realize real-time synchronization and fault rehearsal of a physical entity and a virtual mirror image, and outputting a rehearsal result; constructing a fault prediction model based on an optical network physical mechanism, inputting parameters and rehearsal results to complete feature fusion and trend analysis, and outputting a fault prediction result; an intervention strategy is generated and subjected to twinborn virtual simulation verification; and after the verification is passed, driving the physical layer to execute intervention, and feeding back effect data to dynamically update twin model parameters. A closed loop of perception, modeling, prediction, intervention and feedback is formed, and the problems of prediction lag and low operation and maintenance efficiency in the prior art are solved.

The invention relates to the technical field of pile foundation deformation measurement, in particular to a pile foundation deformation distributed test method and system, and the method comprises the steps: obtaining a time sequence strain value sequence of sampling points of a tower pile foundation, calculating a dynamic disturbance index, extracting a quasi-static signal, and determining a long-term strain value increment through time domain trend analysis; through space-time decoupling operation, correlation between time evolution behaviors of a sampling point and a physically adjacent sampling point is compared, and a space localization index is determined; taking the spatial localization index as a confidence coefficient weight, and performing weighted distribution on the long-term strain value increment to respectively determine an adaptive global trend component and a local abnormal component; and finally, performing spatial integral operation on the global trend component and the local abnormal component to obtain two decoupled deformation curves so as to determine the deformation state. The method improves the accuracy of the deformation test.

The invention belongs to the technical field of physical ability analysis and diagnosis, and particularly relates to a latent variable fusion modeling method and system for old people weakness trajectory prediction based on biomechanics. According to the method, the plantar pressure, the joint angle and the electromyographic signal in the gait process of the elderly are collected through wearable sensing equipment, multi-dimensional structured feature parameters are extracted, and a key index group reflecting the weakness degree is generated based on preset weight fusion. And in combination with time sequence fitting and variation trend analysis, dynamic evaluation and risk level judgment of the weakness state of the old people are realized. The system comprises a data acquisition module, a feature extraction module, an index fusion module, a trend analysis module and a result output module, has the characteristics of high accuracy, strong real-time performance, good interpretability and the like, and is suitable for weakness risk screening and rehabilitation intervention guidance.

The invention discloses a high arch dam global deformation analysis method and system fusing InSAR and measuring point data, particularly relates to the technical field of radar interference deformation monitoring, and is used for solving the problem of inaccurate shielding failure compensation. According to the method, data stability is guaranteed by obtaining InSAR image data and extracting sight line direction displacement in combination with coherence indexes, a key measuring point deformation sequence is collected, an InSAR result is calibrated based on stable reference points, data consistency is improved, a projection correction coefficient is constructed by using a dam body three-dimensional structure model, the sight line displacement is inversely solved into structural normal displacement, and the accuracy of the dam body three-dimensional structure model is improved. Obtaining an initial deformation value of the real structure meaning; according to the method, a spatial coupling map of measuring points and InSAR pixels is constructed for a shielding failure area, deformation compensation of the failure area is completed based on an edge weight, finally the deformation compensation result is remapped to a three-dimensional structure model, measuring point data restoration is performed on a high arch dam, and structural plane-level deformation visualization and trend analysis are realized more quickly and more accurately.

The invention relates to the technical field of building carbon emission management, in particular to a building full-life-cycle carbon emission management system and method based on a calculation management system. The system comprises a data management module for acquiring and processing carbon emission data in a full life cycle stage of a building, mapping the processed carbon emission data with a factor database through a uniform interface and then storing the processed carbon emission data in the database; the carbon emission accounting module is used for executing carbon emission accounting and simulation and uncertainty analysis; the intelligent management module is used for carrying out real-time monitoring, carbon emission trend analysis, abnormity early warning and visual display on the carbon emission data; and the optimization emission reduction module operates an optimization algorithm, generates an emission reduction scheme through iterative processing and performs carbon asset management. According to the building full-life-cycle carbon emission management method based on the calculation, management and reduction system, closed-loop cooperation of accurate accounting, intelligent management and emission reduction optimization is achieved.

The invention discloses an electronic watt-hour meter intelligent data processing system and method based on remote communication, and relates to the technical field of watt-hour meter data processing.The method comprises the steps that windowing management is conducted on electric energy data, data pre-evaluation is conducted at the same time, and data windows needing to be processed are screened out; adjusting infrastructure constraints to reduce the basic capability limitation of decision processing; performing hierarchical judgment on the data window marked as needing to be processed so as to optimize a decision processing level, and performing gap data processing according to a hierarchical judgment result; key data of all electric energy meters and corresponding time periods are collected and analyzed in real time, data portraits of different objects and time periods are constructed, analysis is carried out according to various data anomalies and operation states, various decision parameters are dynamically adjusted, efficient processing and decision optimization of the data of the electric energy meters are achieved, and the data processing efficiency is improved. Therefore, the problem of low accuracy in dynamic statistics and trend analysis of the acquired real-time electric energy data caused by improper decision processing of the data in the prior art is effectively solved.

The invention relates to the technical field of agricultural intellectualization, and discloses a fruit tree growth monitoring system based on multi-sensor fusion, which comprises a multi-mode sensor array, an integrated soil humidity sensor, a three-dimensional ultrasonic ranging sensor, a hyperspectral imaging sensor and a microenvironment meteorological station, and realizes multi-parameter synchronous acquisition; a nonlinear data calibration model adopts a layered architecture, and cross interference between sensors is eliminated; a dynamic weight distribution algorithm is built in the edge computing node, multi-source heterogeneous data are fused in real time, a fruit tree growth situation map is generated, and a low-power-consumption processor is adopted to support lightweight AI model deployment; the self-adaptive power supply module adopts a photovoltaic-temperature difference composite power generation structure to realize cable-free deployment; the wireless ad hoc network communication unit constructs an inter-tree relay network based on a LoRa protocol, supports low-power-consumption remote data return, and integrates a relay node health degree evaluation mechanism. The real-time performance of fruit tree growth monitoring, the data reliability and the situation analysis accuracy are improved.

The invention discloses a production control method and system based on multi-source data fusion and an intelligent algorithm, and relates to the technical field of automatic control. Multi-dimensional production state information is collected through a sensor, multi-source heterogeneous data is integrated into a data set in a unified format through a data fusion technology, and the multi-dimensional production state information is obtained; a long short-term memory network is used for trend analysis and prediction, a multi-objective optimization algorithm is used for calculating an optimal control parameter, multi-loop cooperative adjustment is realized, an adjustment instruction is issued through a distributed control protocol, a priority ranking algorithm is introduced to ensure system stability, and a game theory negotiation algorithm is used for realizing dynamic resource allocation among production units. The production scheduling plan is optimized through the genetic algorithm, continuous monitoring and optimization can be achieved, control parameters can be adjusted in a self-adaptive mode, the production efficiency and the resource utilization rate are effectively improved, and intelligent automatic control under the complex production environment is achieved.

The invention discloses a monitoring data trend analysis and decision suggestion generation method based on a large language model agent. The method comprises the following specific steps: step 1, constructing a domain knowledge enhancement module; 2, developing and registering a callable tool set, and compiling a series of performance functions by using Python; 3, constructing an agent arrangement platform and a core agent, and creating a master control agent in the agent arrangement platform as a scheduling center of an analysis task; step 4, executing an intelligent agent analysis process, and when an analysis request is received, enabling the intelligent agent to autonomously run according to the following process; 5, system integration and interface calling are carried out, the whole agent is packaged into RESTful API service and deployed in an enterprise server, by introducing a large language model agent architecture, an RAG enhanced reasoning mechanism and a callable tool set, the technical transition from passive data display to active insight generation is achieved, and the method has the advantages of being high in practicability and easy to popularize. And the analysis efficiency, the decision accuracy and the system intelligence level are remarkably improved.

The invention discloses an attack trapping and analyzing method and system based on a honeynet, and relates to the technical field of attack analysis, and the method comprises the steps: collecting attack interaction data and an associated attack chain trajectory set; dividing attack behavior units, and constructing a behavior space attitude matrix; constructing a two-dimensional behavior relationship graph, performing interpolation enhancement of sub-time granularity, and raising the dimension of the interpolated behavior node into a three-dimensional semantic space; performing attack path trend analysis on the attack behavior unit, and constructing an attack path bending rule model; and inputting parameters such as the attack chain trajectory set and behavior nodes in the three-dimensional semantic space into the attack path bending rule model, outputting a trapping response strategy, and dynamically adjusting the honeynet environment. According to the invention, the type and position of the honeypot can be automatically adjusted according to the attack behavior change, the structure definition and processing precision of the attack behavior data are effectively improved, and the attack behavior identification result can be conveniently and directly used for trapping strategy optimization.

The invention discloses a multi-modal neural network-driven product style trend analysis method and system, and relates to the technical field of image processing, and the method comprises the steps: receiving product multi-modal data, including text, image and audio data, carrying out the tensor extraction of the product multi-modal data, enabling each modal data to correspond to a single-peak tensor, and carrying out the tensor extraction of the product multi-modal data; carrying out feature fusion on the single-peak tensor to obtain a multi-modal feature tensor, and then carrying out dimension reduction processing; calculating a cross-modal multi-head attention score by using the single-modal tensor based on a language tensor, wherein the cross-modal multi-head attention score is used for weighting single-modal features; fusing weighted features, dimensionality reduction tensors and manually labeled style labels, constructing a supervised training target, and generating single-mode pseudo labels for multi-task model training; cross-modal shared information capture is carried out on the unimodal tensor based on a covariance matrix, cross-modal shared information is constrained through modal alignment loss, and under a multi-task learning framework, a trained multi-task model takes output of a multi-modal main task as a final style label to realize product trend analysis.

The invention discloses an intelligent landslide risk trend analysis method and device based on a trend prediction model, and relates to the technical field of landslide risk data analysis. According to the invention, the micro-core pile and the auxiliary sensor are adopted to collect multi-source time sequence monitoring data, intelligent data analysis processing is carried out based on built-in workflow of the micro-core pile, interference data optimization and data quality analysis are completed, and high reliability of data is ensured; intelligent analysis is carried out based on the trend prediction model, future landslide risks are predicted, and model prediction stability judgment and feedback adjustment are carried out; in combination with real-time monitoring data and known environmental factors, such as rainfall, risk judgment and early warning execution are performed. Through multi-level data optimization, stability feedback adjustment and comprehensive analysis of environmental factors, the precision and real-time performance of landslide risk prediction are effectively improved, the early warning capability is enhanced, and particularly, a more stable and reliable risk analysis result can be provided in a low electric quantity and external interference environment.

The invention provides a heavy-duty truck vehicle-mounted safety monitoring method and device, and the method comprises the steps: obtaining the original monitoring data of a wheel-rail force through a wheel-rail force monitoring module which is installed on an axle and has a low power consumption characteristic, and carrying out the edge calculation of the original monitoring data, and obtaining a wheel load shedding rate, a derailment coefficient and a wheel axle transverse force; the framework acceleration and the bearing temperature are obtained through the low-power-consumption vibration acceleration sensing module and the bearing temperature sensing module; performing multi-modal fusion on each index to obtain a multi-modal operation state representation vector; obtaining a dynamic threshold value of each index based on an enhanced feature vector and a multi-modal operation state representation vector obtained by an offline self-learning model; and giving current situation analysis and trend analysis of safety monitoring based on a comparison result of each index and a dynamic threshold value. According to the invention, the limitation of independent analysis of single sensor data in the prior art is overcome, the dynamic adjustment of the safety threshold is realized, and the accuracy and reliability of safety monitoring are improved.