19668 results about "Data input" patented technology

The invention relates to the technical field of artificial intelligence, in particular to an automatic construction method of an end-to-end agent based on graph structure semantic fusion. The method comprises the following steps: receiving business demand data input by a user; business target and demand constraint condition analysis is carried out on the business demand data, and a core workflow framework of the intelligent agent is generated; performing end-to-end execution path analysis on the core workflow framework of the intelligent agent to obtain an end-to-end workflow; constructing a dynamic evolution semantic map; and constructing an end-to-end call chain execution strategy based on the end-to-end workflow, and performing agent instance packaging and agent instance reinforcement learning enhancement processing according to the dynamic evolution semantic map, thereby automatically constructing an end-to-end agent. According to the invention, by fusing the graph structure knowledge and the generation capability of the large language model, an efficient, accurate and extensible agent automatic construction scheme is provided for various complex business scenes.

The invention provides an intelligent subway equipment fault diagnosis method and system assisted by a large language model, and relates to the technical field of data processing, and the method comprises the steps: extracting key information through a large language model, constructing a multi-dimensional equipment fault knowledge graph, obtaining a historical fault data set, and extracting key fault monitoring parameters related to the fault, obtaining a sensing state data set, and judging whether the sensor state data is abnormal or not; and calling a pre-constructed sensing distortion correction algorithm, generating a fault monitoring correction parameter and executing parameter correction, inputting multi-dimensional monitoring data into a diagnosis engine driven by a large language model, and outputting a most possible fault type, cause analysis and recommendation processing strategy and a fault identification report. The technical problems that in the prior art, due to the lack of fusion modeling capacity for the unstructured fault text and the structured monitoring data, the intelligent degree of fault diagnosis is low, and accurate recognition and causal analysis are difficult to achieve are solved, and the fault recognition response speed and accuracy are improved.

The invention relates to the technical field of equipment fault diagnosis, and discloses an equipment fault diagnosis and prediction method based on deep learning, and the method comprises the following steps: S1, collecting multi-modal data in real time through a plurality of sensors installed on equipment; s2, preprocessing the collected data; s3, constructing a hybrid deep learning model; s4, dynamic weighted fusion is performed on the features of different modal data by using an attention mechanism, and comprehensive feature representation is generated; s5, using the marked fault data and normal data to supervise and train the model; s6, inputting equipment operation data acquired in real time into the trained model, and judging the state of the equipment; and S7, generating a potential fault early warning signal based on a prediction result of the model. A piezoelectric vibration sensor and a thermal infrared imager are arranged on a motor bearing through vibration, temperature and sound sensors, vibration waveforms, thermal imaging slices and time-frequency diagrams are synchronously captured, and composite state characteristics such as mechanical wear and temperature anomaly of equipment are comprehensively reflected.

The invention relates to the technical field of digital twinning, and discloses a pump station working condition monitoring method and system based on digital twinning and a storage medium. The method comprises the following steps: receiving a pump station operation characteristic matrix of pump station equipment through a PLC (Programmable Logic Controller) control cabinet; constructing a hybrid digital twin model based on the pump station operation characteristic matrix and generating operation state prediction data; inputting the operation state prediction data and the actual monitoring data into a Transform network to carry out working condition feature association mode extraction and health index quantitative calculation to obtain an equipment health assessment index; and triggering an early warning response mechanism by using the equipment health assessment index, and outputting a unit start-stop control strategy of the pump station equipment to the scheduling system through the remote management platform. According to the invention, the capability of identifying abnormal working conditions and the accuracy of early warning are improved, the purposes of'few people on duty, remote monitoring and low-consumption operation 'of the pump station are achieved, the operation and maintenance cost is greatly reduced, and the service life of equipment is prolonged.

The invention relates to the technical field of artificial intelligence and natural language processing, in particular to a dynamic question and answer matching method and system based on context awareness. The method comprises the following steps: acquiring history of multiple rounds of dialogues and user background information, and generating a semantic understanding model of a current dialogue; based on the semantic understanding model and the current user input, matching the most relevant answer from the question and answer library, and dynamically adjusting the matching strategy; acquiring real-time feedback data of the user, updating a personalized recommendation algorithm based on the interest change of the user, and generating a personalized answer recommendation list; dynamically updating the content of the question and answer library according to the personalized recommendation list and the user behavior data; and inputting the user feedback and the updated question and answer library data into a personalized recommendation system to generate final personalized answer recommendation. Through context sensing and real-time dynamic adjustment, the matching precision of the question answering system is remarkably improved, the interaction experience of the user is improved, and the adaptability and flexibility of the system are enhanced.

A data processing system comprising: inputting a tiled image of a vehicle including four different angle views of the vehicle combined into a single image to a first machine learning model (e.g. CNN), the model trained based on historical image data to predict a first likelihood of total loss vehicle; inputting a multi-fusion of images each into a second set of machine learning models; the multi-fusion of images including a set of separate and distinct images for each of the views input separately into the second set of machine learning models, and extracting features to predict a second likelihood of total loss vehicle; inputting tabular data relating to the vehicle into a third machine learning model to predict a third likelihood of total loss vehicle for the vehicle; and aggregating the first, second and third likelihood of total loss vehicle to determine the overall likelihood of total loss.

The invention relates to the technical field of unmanned vehicle control, and discloses a smart city environmental sanitation unmanned vehicle path planning and real-time monitoring method. The method comprises the following steps: firstly, acquiring a multi-source environment sensing data set such as laser radar point cloud data, a camera image sequence and real-time traffic flow information; local road network features are extracted based on laser radar point cloud data, a dynamic target motion prediction map is generated according to a camera image sequence, and real-time traffic flow information is processed to generate a regional traffic efficiency evolution curve. And inputting the data into a path optimization model to generate an initial path sequence, dividing cleaning task priorities, fusing related data and generating a final path planning scheme through a reinforcement learning algorithm. In addition, operation state data of the unmanned vehicle are collected in real time, and a path correction instruction set is generated through an anomaly detection model to update the strategy network. According to the method, the rationality and the operation efficiency of the path planning of the environmental sanitation unmanned vehicle can be improved, real-time monitoring is realized, and the operation safety and the management intelligence level are enhanced.

The invention discloses a dynamic visual content collaborative generation method based on a cognitive logic chain, and belongs to the technical field of visual content generation, and the method comprises the following steps: S1, user intention analysis and data input; s2, dynamically constructing a cognitive logic chain; s3, intelligent scheduling of the multi-modal generation module; s4, cross-modal content collaborative generation is carried out; s5, collaborative editing and real-time feedback are carried out; s6, iterative optimization of logic chain driving; s7, multi-dimensional quality evaluation: constructing an evaluation matrix containing semantic consistency, visual attraction and user participation degree, predicting a content propagation effect in combination with a deep learning model, and generating a quantitative improvement suggestion report; and S8, updating the self-adaptive knowledge reversely marking the cognitive logic chain according to the finally adopted content version, extracting a new association rule, and injecting the new association rule into the rule base. Through deep semantic analysis and dynamic logic chain construction, the system accurately captures a core creation target of a user and converts the core creation target into an executable visual strategy.

The invention discloses a ppb-grade methane gas online detection method, equipment, a medium and a product, and relates to the field of data processing. The method comprises the following steps: acquiring an infrared absorption spectrum signal, and obtaining original spectrum data; performing multi-scale noise separation and baseline correction on the original spectral data to obtain de-noised effective spectral features, and extracting methane specific absorption peaks from the de-noised effective spectral features; comparing the methane specific absorption peak with the interference gas spectral line through a multi-spectral line weight matching algorithm to obtain methane characteristic intensity; carrying out self-adaptive calibration on the methane characteristic intensity according to the drift compensation model to obtain calibrated characteristic intensity; obtaining a ppb-level methane concentration value through a concentration inversion model; and inputting the sorted concentration time sequence data into an anomaly detection model to detect a methane leakage event. By designing signal processing and algorithm analysis, the dependence on equipment hardware performance is reduced, and the use cost of ppb-level methane gas detection is reduced.

The present application relates to the field of wire laying control technology, and discloses a composite material automatic wire laying forming optimization method, system, medium and wire laying machine. The method comprises: inputting dynamic tension control data into a fuzzy neural network calculation engine to obtain a dynamic compensation model for the retraction and release speed; based on the dynamic compensation model for the retraction and release speed, using a reinforcement learning engine to calculate the shearing sequence of multiple bundles and generate an intelligent shearing instruction set; constructing a re-delivery trajectory mapping relationship based on the intelligent shearing instruction set, and forming a re-delivery path matrix after iterative calculation by a digital twin engine; establishing a temperature field distribution prediction model based on the re-delivery path matrix, and constructing a temperature field dynamic control sequence after finite element analysis and calculation; substituting the temperature field dynamic control sequence into a multi-objective constraint equation, and outputting a laying strategy data packet after deep reinforcement learning iterative optimization. The present application realizes the precise coordination of multiple process parameters such as tension control, speed compensation, shear control, and temperature control.

The invention provides an audit report automatic generation method based on natural language processing, and relates to the technical field of natural languages, and the method comprises the steps: obtaining audit document data passing the examination and approval, the audit document data comprising an audit manuscript, an audit problem, an audit evidence obtaining sheet and an audit verification sheet; performing semantic analysis on the audit document data through a natural language processing model, and extracting key information of an audit subject, a risk point and a compliance conclusion; inputting the key information into a structured processing algorithm to generate structured data matched with a preset auditing template; and inputting the structured data into a pre-training text generation model for language reconstruction, generating an audit report text, automatically filling the audit report text into a corresponding field of an audit template, and outputting a complete audit report.

The invention relates to the field of frequency converter fault detection, and discloses a frequency converter fault prediction method and system based on machine learning, and the method comprises the steps: obtaining multi-dimensional real-time data in the operation process of a frequency converter; constructing a dynamic mapping relation to obtain a basic feature set; generating a time sequence feature vector capable of reflecting the state change of the equipment based on the basic feature set; comparing, analyzing and judging whether the equipment state deviates from a normal operation interval or not based on the historical operation data and the time sequence feature vector, and outputting a state deviation index; performing abnormal fluctuation judgment on the time sequence feature vector; extracting fluctuation amplitude and frequency characteristics of the key indexes to obtain quantitative description data of abnormal fluctuation; inputting the quantitative description data of the abnormal fluctuation into an abnormal prediction model; and generating a coping strategy and a triggering condition of the coping strategy based on the risk prediction result. The method has the advantages that the abnormal state of the frequency converter is recognized in time, and potential risks are predicted.

The invention relates to the technical field of network security, in particular to a computer network information security monitoring method, system and device and a medium, and the method comprises the steps: obtaining encrypted traffic data and application log data in a network environment, and constructing a space-time associated original data set based on the encrypted traffic data and the application log data; performing protocol analysis on the original data set with the time-space association to generate a protocol fingerprint feature vector; inputting the protocol fingerprint feature vector and the application log data into a preset heterogeneous multi-modal analysis model, and obtaining a multi-dimensional security situation assessment result containing a threat level and an attack path; and based on the multi-dimensional security situation assessment result, generating a dynamic defense strategy instruction set through a reinforcement learning algorithm, and issuing a strategy instruction to a network execution node in real time. The method and the device have the effects of realizing real-time accurate detection of encryption threats and constructing a dynamic defense system with balanced security and efficiency.

The invention provides an optical energy storage regulation and control method and system based on multi-time scale rolling optimization in the technical field of energy storage regulation and control, and the method comprises the steps: S1, inputting a 24-hour electricity price prediction curve, a 24-hour photovoltaic output prediction curve, a 24-hour load prediction curve and an initial SOC into a day-ahead prediction model, a 24-hour SOC reference trajectory and a standard charging and discharging plan are obtained; s2, collecting feedback data, and inputting the 24-hour SOC reference trajectory, the standard charging and discharging plan and the feedback data into the intra-day prediction model to obtain a corrected charging and discharging plan; and S3, setting an SOC dynamic limit value, collecting a real-time electricity price and real-time photovoltaic output, and inputting the corrected charging and discharging plan, the SOC dynamic limit value, the real-time electricity price, the real-time photovoltaic output and the actual SOC into the real-time prediction model to obtain a charging and discharging instruction and frequency modulation output power. The method has the advantages that the precision, economical efficiency and power grid adaptability of energy storage regulation and control are greatly improved, and the attenuation speed of the service life of the battery is greatly delayed.

The invention provides a shield intelligent auxiliary type selection system and method based on a large language model. The model selection system comprises a data input module, a rule knowledge base module, a large model reasoning module and a result generation module. The integrated decision-making system integrating a rule knowledge base, a deep learning model and expert system logic is constructed for the practical problems of complicated geological conditions, multiple rule constraints, high expert dependency and the like in shield construction, and the system combines a structured model selection rule and historical case data, and has the advantages of intelligence, standardization, self-learning, high efficiency and the like. The problems of low efficiency, high subjectivity, insufficient intelligent degree and the like of the existing shield tunneling machine model selection depending on artificial experience and partial standardized guide are solved, and the transformation of shield construction management from artificial experience to intelligent decision can be promoted.

The invention belongs to the technical field of power distribution network voltage regulation and control, and discloses a distributed photovoltaic complex power prediction-cluster-cooperation-based power distribution network voltage regulation and control method, which integrates photovoltaic historical data, inputs an improved back propagation neural network model and outputs predicted photovoltaic active power output. Estimating the reactive capacity boundary of each node in real time based on the running state of the network-following inverter; dividing a distributed photovoltaic cluster by establishing a two-dimensional modularity function of a net load index and an equivalent electrical distance; a multi-device differential cooperative control strategy is provided for the voltage out-of-limit risk in the cluster; and constructing an optimization function with minimum network loss and voltage offset as a target, and optimizing and solving the function by using an improved multi-organization particle swarm optimization algorithm to obtain a multi-device adjustment sequence and a device action amount. According to the method, the renewable energy consumption capacity is improved and the network loss is reduced while the voltage stability of the power distribution network is ensured, and the comprehensive adjustment cost is optimized.

The invention discloses an electric power system abnormal remote signaling detection method based on a graph auto-encoder model, and the method comprises the steps: collecting measurement data of an electric power system, carrying out the preprocessing, obtaining a graph data set with an abnormal label, and dividing the graph data set into a training set and a test set; inputting the graph data in the training set into the graph auto-encoder model for training; after training is completed, abnormal score distribution is counted based on normal edge samples in a training set, a threshold value is set to serve as a follow-up judgment basis, and threshold value selection takes the accuracy rate and the recall rate on a test set as an adjustment and optimization target; in a test stage, image data in a test set are input to carry out edge feature reconstruction and anomaly scoring, anomaly judgment is carried out on edges in combination with a set threshold value, and a preliminary abnormal edge detection result is output; and the output abnormal edge detection result is input into the graph restoration module, the restored edge structure and edge features are output, the damaged remote signaling state in the power grid is restored, and the integrity of the graph structure and the operation credibility of the power system are improved.

The invention relates to the technical field of adaptive control, in particular to an unmanned aerial vehicle adaptive flight control system and method based on multi-modal fusion, and the system comprises an airspace analysis module, a signal screening control module, a state recognition module, a parameter matching module and a feedback adjustment module. According to the method, through a multi-source task data analysis and trend signal screening mechanism, the accuracy and real-time performance of path planning and adjustment are improved, through combination of channel fluctuation trend identification and abnormal state locking, the stability of sensing data input is enhanced, and the identification accuracy of an unbalance state is analyzed and optimized by using attitude change direction consistency. The attitude error trend and control parameter segment similarity comparison mode is adopted, parameter matching has dynamic responsiveness, control weight configuration is adjusted through collaborative analysis of the path direction component and the attitude adjustment trend, rhythm self-adaptive adjustment of aircraft control output is achieved, and the control precision of the aircraft is improved. And the dynamic adjustment capability and the safety redundancy response capability of flight control are enhanced.

The invention provides a motion control method and system for an intelligent robot, and the method comprises the steps: collecting environment and state data through an intelligent sensor group of a humanoid robot, inputting the environment and state data into a pre-training first neural network model, and obtaining motion prediction data and an environment analysis result; constructing a motion planning model, and performing energy consumption-stability multi-objective optimization on the joint motion track by adopting a preset first algorithm; the central controller generates a joint position, speed and torque reference trajectory based on an optimization result; and the local second controller of each joint locally adjusts the reference trajectory within the prediction time domain according to the real-time feedback. According to the invention, multiple sensors are combined with the mixed attention neural network to realize environment and self state intelligent perception, and the problem of multi-sensor data fusion time sequence dependence is solved; through energy consumption-stability multi-objective optimization, the complex environment movement efficiency is remarkably improved; the central controller and the local controller work cooperatively, and in combination with an edge computing architecture, the communication delay is reduced, and the system response speed is improved.

The invention discloses an intelligent water and fertilizer integrated precise irrigation control method. The method comprises the following steps: step 1, establishing a multi-modal data acquisition network; step 2, inputting the data obtained in the step 1 into a trained LSTM neural network, and predicting future crop transpiration water demand; 3, according to nutrient concentration distribution fed back by a soil conductivity sensor in real time, combining with a water and fertilizer coupling migration model, dynamically calculating a nitrogen-phosphorus-potassium ratio of each irrigation zone, realizing precise mixing of a fertilizer mother solution and irrigation water through a three-channel proportioning pump, and implementing a variable irrigation decision; 4, dynamically adjusting the pulse irrigation duration of each branch through an MPC algorithm; and step 5, recovering the residual liquid at the tail end of the pipe network to a buffer tank, detecting the EC value of the recovered liquid through an electrochemical sensor, and taking the EC value as a reference parameter automatic correction proportioning algorithm of the next irrigation period. Accurate control over the irrigation process can be achieved, so that the water and fertilizer utilization efficiency is improved, and resource waste and environmental pollution are reduced.

The invention provides an energy storage battery health state evaluation and thermal runaway risk prediction method and system, and belongs to the technical field of energy storage batteries. The method comprises the following steps: establishing an electrochemical fusion model by fusing an electrochemical model, a thermal model and an aging model; training the electrochemical fusion model, and inputting physical parameter data of the energy storage battery into the trained electrochemical fusion model to obtain key performance parameter data of the energy storage battery; obtaining an energy storage battery health state evaluation result based on the key performance parameter data of the energy storage battery; and comparing the key performance parameter data of the energy storage battery with data detected by an actual sensor to obtain a thermal runaway risk prediction result. Accurate evaluation and fault prediction of the energy storage battery health state are achieved, the precision and safety of battery management in an energy storage system can be remarkably improved, and the method is suitable for a large-scale energy storage system.

The provided are a price elasticity analysis and prediction method and model based on deep learning. The model consists of a CNN layer and an RNN layer. The method comprises the following steps: S1, collecting historical data and merging the historical data into a multi-dimensional time series dataset; S2, extracting sentiment data and trend data from market news and social media; S3, inputting the data obtained into CNN for data preprocessing and feature extraction; S4, inputting the feature extracted by CNN into RNN for time series analysis; S5, training and optimizing model: using Adam algorithm to adjust the learning rate adaptively, and combining the momentum method and RMSProp algorithm to improve the generalization ability and prediction accuracy of the model. The provided combines the advantages of CNN and RNN, which can understand and predict the complex relationship between price and market behavior more comprehensively and accurately.

The invention discloses a fault detection and root cause analysis system for a multi-source energy system, and relates to the technical field of energy management and fault diagnosis. The system obtains various operation data and image data of the multi-source energy system in real time; preprocessing the operation data and the image data; inputting the preprocessed data into a hybrid algorithm combining wavelet transform, CNN and LSTM, and detecting an abnormal mode and a potential fault in the multi-source energy system; the digital twinborn model is dynamically updated according to data acquired in real time; and in combination with the feature extraction capability of a digital twin model and a hybrid algorithm, positioning key causes causing the fault, simulating a system operation state, analyzing potential causes of the fault, and outputting an analysis result. According to the invention, the comprehensive analysis capability of the system on various data sources is effectively improved, and a comprehensive and accurate solution is provided for intelligent operation and maintenance of the energy system.

The invention relates to an underwater image enhancement method based on frequency domain analysis and visual Mama, comprising the following steps: constructing a multistage wavelet transform network, the network being an encoder-decoder structure, the encoder comprising a space-frequency feature fusion module and discrete wavelet transform; the decoder comprises a space-frequency feature fusion module and inverse discrete wavelet transform; an underwater image is decomposed into a low-frequency part and a high-frequency part through discrete wavelet transform in an encoder, spatial feature extraction and frequency feature extraction are conducted on the low-frequency part and the high-frequency part through corresponding space-frequency feature fusion modules based on Mama and frequency domain convolution respectively, and splicing is conducted on the low-frequency part and the high-frequency part along a channel in the decoder according to the sequence of low frequency and high frequency. Reconstruction is carried out through inverse discrete wavelet transform; and inputting the underwater image data into the multistage wavelet transform network to obtain enhanced underwater image data. Compared with the prior art, the method has the advantages of improving the enhancement quality, reducing the complexity, quickly enhancing and the like.

The invention relates to an event information enhancement and deduction prediction method combined with a knowledge graph, and belongs to the field of natural language processing and knowledge graphs. The objective of the invention is to solve the problem of insufficient accuracy and robustness of a deduction result due to the fact that multi-dimensional information cannot be effectively fused in an existing event deduction method. In order to solve the problems that a traditional event deduction method mostly depends on a large amount of entity link information and is low in efficiency and prone to being influenced by data scarcity in practical application, the method comprises the five steps of data input and preprocessing, knowledge graph construction and optimization, event semantic background enhancement, event deduction prediction and event deduction prediction. According to the method, the limitation of a traditional method in event deduction can be effectively overcome by introducing technologies such as a cross-domain knowledge graph and a graph neural network, rich event semantic backgrounds are constructed by utilizing the knowledge graph, event information is enhanced and deduced in combination with a time sequence and a causal relationship between events, and the event deduction efficiency is improved. Therefore, the accuracy and reliability of event deduction are improved.

The invention discloses an intelligent vehicle extreme test scene generation method based on deep learning. The method comprises the following steps: building a hybrid generative adversarial network scene; collecting and preprocessing data; performing dynamic confrontation training; according to the extreme test simulation, an intelligent vehicle extreme test scene generation system based on deep learning comprises a hybrid generative adversarial network model, a data input module, a preprocessing module, an edge cloud module and an interaction module. A double-channel confrontation mechanism enables the risk controllability of a generated scene to be high, and dynamic confrontation training is carried out; a real multi-source data set is enhanced and labeled, the integrity of point cloud data is high through a point cloud missing region feature completion algorithm, and dynamic risk entry is effectively simulated by introducing a social force model into dynamic obstacle modeling; the multi-modal risk assessment network improves the assessment accuracy.

The invention relates to the technical field of mode recognition, in particular to a transformer substation equipment temperature measurement method and system based on visible light guidance. Firstly, a visible light image, an infrared thermogram and vibration spectrum data are synchronously collected, a dynamic compensation coefficient matrix is constructed through illumination intensity analysis, the dynamic compensation coefficient matrix is applied to the infrared thermogram, and illumination interference compensation is achieved. And then, inputting the corrected infrared thermogram and visible light image into a pre-trained heterogeneous atlas matching model, and identifying the category and three-dimensional coordinates of the target equipment. And based on the equipment category information, calling a standard temperature field template in the digital twin model library, establishing a heat conduction path model in combination with the vibration spectrum data, and generating a temperature field prediction reference map of the equipment. And finally, through pixel-by-pixel deviation analysis, inputting the temperature deviation value, the image features and the vibration data into the probability graphic model, and outputting equipment fault information. The method can realize high-precision temperature measurement and fault diagnosis.

The invention belongs to the technical field of electromechanical equipment intelligent monitoring and operation and maintenance management, and discloses an electromechanical system health monitoring and optimization method based on artificial intelligence, and the method comprises the steps: building a digital twin architecture based on a simulation model, reasonably arranging a multi-source sensor, and carrying out the optimization of the health of an electromechanical system; key equipment is fully covered; the system acquires operation data at high frequency and synchronously transmits the operation data to the cloud through multiple channels to realize real-time monitoring; the collected data is cleaned and preprocessed to remove abnormities so as to ensure the data quality; inputting the cleaned data into a neural network, extracting features, and identifying an operation trend and a fault mode; establishing a prediction model in combination with historical data, and constructing a maintenance knowledge base; the equipment state is predicted in real time, and early warning is dynamically generated; through virtual-real synchronization, health and carbon emission states are mapped in real time, equipment operation is intelligently optimized, and cloud closed-loop monitoring is continuously iterated.

The invention discloses a double-flow remote sensing image change detection method fused with Mama enhancement, which belongs to the field of image processing and comprises the following steps: acquiring images to obtain a remote sensing image data set; preprocessing the obtained dual-time-phase remote sensing image, and dividing the image into a training set and a test set; designing a dual-flow change detection network model integrated with Mama enhancement; training the constructed double-flow change detection network model by adopting training set data until the whole model is converged, and storing an optimal model; and inputting test set data into the trained optimal model, and predicting a change area in the test set. According to the method, MambaBlock and a semantic segmentation aggregation module are introduced, the global feature modeling capability is enhanced under linear complexity, and the detection precision and robustness of a change region are improved by fusing feature information of different levels and multi-scale feature learning.

The present invention relates to the field of ship design, and in particular to an automatic three-dimensional modeling method, system and device based on a two-dimensional ship CAD drawing. The method includes obtaining and parsing ship CAD drawing data, extracting metadata from the drawing, and identifying primitives; generating images through rendering technology for unidentified complex primitive areas, and performing preprocessing; inputting the preprocessed images into an object detection model, generating detection frames and category annotations using a deep learning algorithm, and performing coordinate conversion to accurately calibrate the positions of primitives; extracting geometric information, structural attributes and material data based on object detection results and primitive information, performing parametric expression conversion, and generating parametric model data of the ship structure; inputting the generated parametric data into a modeling instruction template, generating a ship CAD modeling instruction sequence, and executing the instructions through a CAD system to automatically construct a three-dimensional model of the ship. The present invention can realize the fully automated conversion from two-dimensional CAD drawings to three-dimensional modeling, thereby improving the efficiency and accuracy of ship design.