251 results about "Path recognition" patented technology

The invention discloses a method and a system for processing route planning data for electric power line inspection. The method comprises the following steps: acquiring line distribution and topographic feature data; gridding the power network region, determining the patrol priority, and generating a region priority sequence; the overall patrol task is decomposed, a plurality of task subsets are generated, one task is allocated to each unmanned aerial vehicle, and each task subset comprises an initial position and a target line section; generating an initial route path based on the topographic features; comprehensively analyzing all initial paths, identifying potential airspace intersections and performing conflict prediction, and generating a conflict-free route path set by performing height hierarchical adjustment on related paths; and calculating predicted flight time and energy consumption of each unmanned aerial vehicle, redistributing boundary tasks according to a load balancing principle, and finally generating a balanced multi-unmanned aerial vehicle cooperative route scheme. According to the method, the safe and efficient multi-machine collaborative inspection route can be automatically generated through systematic task decomposition, conflict prediction and load balancing optimization.

The invention relates to the technical field of structure health monitoring, in particular to an immovable cultural relic structure health detection system based on multi-source sensor fusion, which comprises a multi-source texture response acquisition module, a curved surface drift path identification module, a crack region extraction module, a stress response trend offset module and a health level region output module. According to the method, the continuous coding of the structure texture is combined with the micro-strain response characteristics, so that the accurate judgment of the synchronous drift of the surface texture of the cultural relic structure is realized, the identification sensitivity of potential micro deformation is improved, a multi-period track set is constructed according to the center deviation trend in a time sequence, and the evolution tracking capability of space behaviors is enhanced; the crack linkage area is subjected to consistency screening through image gray abrupt change and strain direction, the accuracy of crack boundary identification is improved, the safety of an immovable cultural relic structure is guaranteed, and meanwhile collaborative improvement of real-time monitoring, the response speed and the automation level is considered.

The invention relates to the field of e-commerce recommendation, and discloses a user behavior data analysis system and method based on a time series, and the method comprises the steps: continuously capturing user operation events, extracting a short-term interest fluctuation factor, a category offset index and an interaction strength parameter, and forming dynamic weight mapping for constructing a transient preference map; implicit association detection of a cross-category event chain is executed on the transient preference map, and a multi-dimensional influence gradient vector reflecting the potential interest evolution trend is generated; tracking an offset path of a key interest node through a hierarchical interest-driven algorithm by utilizing a multi-dimensional influence gradient vector, and identifying candidate behavior fragments causing recommendation delay; performing priority label reconstruction on the candidate behavior fragments to form a local behavior network with a weight gradient; based on a local behavior network, key nodes with unbalanced interest distribution in a specific time window are detected, and an evolution path for real-time adjustment of a recommendation engine is generated. The method has the advantage of improving the relevance and instantaneity of the recommended content.

The invention relates to an industrial internet multilayer causal motif abnormal propagation path identification method and system, and the method comprises the steps: firstly carrying out the construction and extraction of a multilayer high-order motif, extracting a motif unit which expresses the local high-order structure features through the construction of a semantic hierarchical graph structure in combination with a frequent sub-graph mining and cross-layer motif alignment mechanism, and carrying out the recognition of the abnormal propagation path of the multilayer causal motif. Stable and uniform multi-layer motif representation is formed; then, on the basis of the structural equation model, motif variables are regarded as endogenous variables of a causal model, a causal path between motifs is mined by introducing conditional mutual information and a Bayesian structure learning algorithm, an average causal effect is calculated to construct a causal consistency matrix, and causal community division is realized in combination with a weighted modularity optimization method; and finally, quantifying the dynamic change of a community causal structure by constructing a causal deviation graph between an expected causal graph and an observed causal graph, and assisting in identifying a causal-driven abnormal propagation path. According to the method and the system, accurate detection and causal traceability of equipment-level and subsystem-level abnormal modes in an industrial system can be realized.

The invention provides an intelligent mine management and control method and system based on the Internet of Things technology, and relates to the technical field of mine management and control, and the method comprises the steps: obtaining the roadway feature data and mining progress data of a mine in a mine; the influence degree on sensor signal transmission is calculated according to geological stress change data and equipment electromagnetic radiation intensity, and the distribution positions of the sensors are planned by combining roadway distribution data. And generating a sensor migration path plan, adjusting a point distribution position to obtain an optimized point distribution position, simulating a signal propagation path by adopting a Monte Carlo algorithm, and identifying a key area of which signal attenuation exceeds a preset threshold value. And adjusting the optimized point distribution position of the key area to obtain key area distribution points, and formulating a sensor distribution and control scheme in combination with the optimized point distribution position. According to the invention, the distribution positions of the sensors are optimized, the monitoring efficiency and the stability of signal transmission are improved, and the real-time monitoring and dynamic adjustment of the mine environment are realized.

The invention relates to the technical field of image recognition, in particular to a blind sidewalk recognition method based on image processing. The method comprises the following steps: collecting a ground scene image, and carrying out noise suppression and illumination normalization processing to obtain a ground scene image to be processed; inputting the to-be-processed ground scene image into a pre-constructed convolutional neural network model to identify feature textures of the blind sidewalk bricks; dividing a blind sidewalk area in the ground scene image to be processed by using the blind sidewalk brick feature texture, and determining a direction gradient feature of the blind sidewalk area; executing blind sidewalk direction consistency constraint based on the direction gradient features to collect blind sidewalk structured path segments; recognizing a blind sidewalk fracture area based on directional gradient features; and performing cross-frame target tracking according to the structured path segment of the blind sidewalk, and outputting a continuous blind sidewalk trajectory. The automatic recognition rate of the blind sidewalk area is improved based on the image recognition technology, and the continuous recognition capacity of the blind sidewalk path in the complex illumination and shielding environment is enhanced.

The invention discloses a clock domain crossing circuit timing sequence closed-loop optimization method and system, and relates to the technical field of integrated circuit design, and the method comprises the steps: reading a circuit netlist and timing sequence constraints, recognizing a converged path, and carrying out the clustering analysis to form a plurality of path groups; establishing a time sequence analysis model, and dynamically calculating the maximum delay constraint of the data signal in the path group; matching and executing at least one time sequence optimization strategy for each path group; and carrying out closed-loop time sequence verification, retaining a legacy path set which does not meet the constraint, and repeatedly executing the time sequence optimization strategy until a convergence condition is reached or the maximum preset number of iterations is reached. According to the method disclosed by the invention, the cross-domain path needing to be optimized in the DMUX structure can be accurately positioned without manual intervention, so that the path identification efficiency is improved, constraint omission or grouping deviation caused by manual operation is avoided, design defects are reduced from the source, and the design reliability of a chip is improved.

The invention provides a power grid fault propagation path identification method and system based on a graph neural network, and relates to the technical field of power grid fault identification, and the method comprises the steps: obtaining a power grid topological structure and fault observation data, constructing a direction perception graph neural network, achieving the one-way message passing based on potential energy difference, and obtaining a node potential energy value and embedded representation; edges are screened to generate a directed graph, and candidate paths are generated in the potential energy gradient direction; and verifying the satisfaction degree of the graph neural network evaluation path to the physical constraint through an invariant, and finally selecting the path with the minimum topological entropy. According to the invention, the accuracy and reliability of fault propagation path identification are improved.

The invention relates to the technical field of acoustic analysis, in particular to a self-adaptive acoustic analysis method and system, and the method comprises the following steps: obtaining a sound pressure amplitude screening channel, extracting a main sound point time index, judging a fluctuation direction to generate a path identifier, and completing channel alignment and path binding to generate an analysis result. According to the invention, by acquiring the maximum sound pressure amplitude of each channel and dynamically screening the effective channel, the capability of capturing significant sound in a complex sound source environment is improved, and by combining directional extraction and fluctuation trend judgment of main sound point time index differences in a continuous period, dynamic perception and track identification of sound source feature changes are realized. The stability of path identification is enhanced by using direction consistency and time sequence alignment operation, accurate binding and real-time adaptive sound source tracking of multiple channels are completed, the identification accuracy and path continuity in a sound source dynamic change scene are effectively improved, meanwhile, redundant channel interference is avoided, and the accuracy of path identification is improved. And the processing efficiency and the adaptive capacity of the acoustic model are enhanced.

The invention relates to the technical field of network security, in particular to a network security operation and maintenance automatic analysis and response system based on artificial intelligence, which comprises an operation sensing module for acquiring a server state, extracting a security log, monitoring a network protocol and a transmission direction, identifying a performance bottleneck, sorting event data and generating a sensing summary sheet, the path offset identification module tracks a cross-domain path, identifies an abnormal path and generates an offset list; the behavior chain construction module sorts an event sequence and generates a behavior chain graph; the response processing module arranges processing steps, records a response process, generates an operation and maintenance execution sequence, synthesizes an output module to complete a process and generates an analysis result. According to the invention, by comparing the server resources with the historical reference, the abnormity is identified, and the event behavior basis is provided. And tracking cross-domain path lag and abnormity, and positioning threats. Arranging event data, and constructing an attack chain. The automatic response improves the speed and accuracy, shortens the response time, and enhances the attack handling capability.

The present application relates to the technical field of supply chain management, and in particular to a supply chain resilience path identification method based on fuzzy set qualitative comparative analysis. The method comprises: acquiring historical operation data of nodes in a supply chain and response data under a sudden incident; on the basis of the similarity of the historical operation data between the nodes, obtaining the degree of operational correlation between the nodes; grouping all the nodes to obtain a plurality of node group clusters; on the basis of the response consistency of the nodes under the sudden incident, screening out a key node group cluster; on the basis of the distribution characteristics of the response data of different nodes, obtaining the response priority of each node; on the basis of the change difference of the response data between each node in the key node group cluster and other nodes in the same batch, obtaining the degree of external interference of each node; and on the basis of the degree of external interference, correcting the response priority to obtain a corrected response priority. The present invention can improve the evaluation efficiency and decision support capability of supply chain resilience.

The invention provides a database fault root cause positioning method and device based on causal discovery, and belongs to the technical field of database operation and maintenance and fault diagnosis. Comprising the following steps: generating a two-dimensional data table and a statistical information set by using a multi-source operation log of a database system; constructing a causal discovery algorithm set A, and training the gradient boosting tree model by using the training set and the path combination to obtain an agent model Fb; establishing a Monte Carlo tree, selecting child nodes of root nodes according to performance expectation and exploration rewards, and expanding the child nodes to leaf nodes layer by layer; running the Monte Carlo tree, and obtaining a causal graph G * and a performance score by using a voting mode according to all causal relationships in the causal graph obtained by each node; calculating an exploration reward and a performance expectation of each node in a complete algorithm path, and returning the exploration reward and the performance expectation upwards to a root node from a leaf node along the path; updating Fb based on the searched path; and processing the new task by using a Monte Carlo tree, and identifying an affected processing variable according to a path pointing to an abnormal result variable in the causal graph.

The invention relates to the technical field of software security, and discloses a variational inference-based micropatch repair path adaptive optimization method, system, device and medium, and the method comprises the steps: constructing a joint control flow diagram and a data flow diagram of a to-be-analyzed program, and forming a program structure diagram; based on the program structure diagram, modeling and sampling potential variables of the repair paths through variational inference, and generating a plurality of candidate repair paths; for each candidate repair path, generating a micropatch statement according to context semantics of each node in the path, and inserting the micropatch statement into a corresponding instrumentation position in the source code; and performing multi-dimensional scoring and sorting on the candidate repair paths, and selecting an optimal repair path for deployment according to a scoring result. According to the method, the structural breakthrough of the whole process is realized in the aspects of path identification, patch generation, strategy optimization, engineering deployment and the like, and a technical support is provided for constructing a software vulnerability repair system with autonomous generation, semantic understanding and controllable deployment capabilities.

The invention discloses a wind generating set fault processing method and device based on multi-modal information, equipment and a medium, and relates to the technical field of wind power generation, and the method comprises the steps: preprocessing the multi-modal fault related information of a wind generating set, and constructing a multi-modal fault database; constructing a multi-modal joint semantic embedding space based on the multi-modal fault database, and determining an optimized current multi-modal large model in combination with a cross-modal hierarchical attention mechanism; dynamically constructing a three-dimensional knowledge graph based on a multi-modal joint semantic embedding space, and determining a path recognition result for a newly occurring target fan fault event in a reinforcement learning process of performing fault propagation path recognition by using the knowledge graph; and determining a structured decision report based on the path identification result and the current multi-modal large model to trigger equipment maintenance operation, and determining a correction feedback report based on the corresponding maintenance record and the current multi-modal large model. Defects in an existing solution can be solved, and the accuracy of fault processing decision is improved.

The invention relates to a low-voltage distribution line comprehensive intelligent protection method based on multi-parameter linkage, and the method comprises the following steps: collecting power distribution abnormal data, calculating a trigger interval, screening segments in a period, generating a sequential segment group, analyzing a serial number difference value to form a continuous path, recognizing rhythm abnormality, and eliminating interference flow. And a time fault is checked to construct a closed chain table, a rhythm stable closed loop structure is screened, and an intelligent protection result is output. According to the method, continuous fragments are constructed based on trigger rhythms and event numbers in abnormal data processing, path continuity is judged through difference values, effective flow directions are recognized, disturbance fragments are eliminated in combination with rhythm comparison, stable closed paths are constructed, and sequences with consistent rhythms and continuous numbers are screened to serve as a judgment basis. Accurate description of an abnormal propagation path and effective tracking of a response rhythm are realized, and identification sensitivity and protection strategy adaptability of a distribution line under multi-source interference are enhanced.

The invention provides a discrete mixed job production line scheduling method based on an improved genetic algorithm, and relates to the technical field of industrial automation and production scheduling, and the method comprises the steps: S1, carrying out the modeling and data input of a production line scheduling problem, constructing a flexible job shop scheduling model, defining a decision variable, a constraint condition and a target function, and inputting basic data; s2, generating an initial population by adopting a hybrid initialization strategy, wherein the initial population comprises randomly generated individuals and individuals generated based on a heuristic rule; s3, chromosome coding is carried out on the scheduling scheme in a two-segment coding mode, wherein a process sorting segment and a machine distribution segment are included; s4, the fitness is calculated, and individual selection is carried out by adopting a tournament selection method; s5, executing improved genetic operations including adaptive crossover and mutation operations; s6, carrying out local search on the elite individuals, wherein the local search comprises key path identification and neighborhood disturbance; and S7, judging a termination condition, if the termination condition is met, outputting an optimal scheduling scheme, otherwise, returning to the step S4.

The invention discloses an intelligent fatigue prediction method and system based on recreation facilities, and belongs to the technical field of intelligent operation and maintaining.The method comprises the steps that stress-strain data, environmental parameters and acoustic emission signals of a target recreation facility in multiple operation modes are collected, and an original multi-source data set is constructed; identifying microcrack initiation time nodes and fatigue early-stage features thereof based on change point detection and spectral analysis; a fatigue response sensitivity matrix is generated by constructing a microscopic response model fused with environmental factors; predicting a microcrack propagation path by using a time recursion model, identifying a fatigue risk point, and automatically generating a maintenance suggestion; if a frequency domain abnormal signal is found, carrying out residual life reevaluation and dynamically optimizing model parameters; according to the method, early fatigue recognition, trend prediction and adaptive optimization can be realized, and the safety and maintenance efficiency of recreation facilities are improved.

The invention relates to the technical field of purchase matching, in particular to an online purchase processing method and system oriented to multiple subjects and configurable rules, and the method comprises the following steps: obtaining parameter contents, verifying semantic relationships, field formats and numerical boundaries, embedding state paths, synchronously updating parameter positions, constructing a rule configuration set, and establishing a rule matching set; dividing a role permission supplement operation path, mapping a multi-subject behavior structure, screening a same-value field pushing round sequence, marking time, and matching a response sequence to judge a path inheritance relationship. According to the method, through a parameter checking and semantic analysis mechanism, the consistency of content embedding paths is kept, a multi-role action sequence is constructed based on permission mapping, an operation path is divided and cooperatively controlled, a field aggregation and timestamp extension rule is identified, a quotation round sequence is constructed, and an action path is inherited and promoted by depending on a response matching judgment mechanism. The state link dynamic state is updated, the interface is synchronized, and the overall linkage of parameter analysis, path control, behavior recognition and process scheduling is enhanced.

The invention discloses a scene-driven multivariate main body data collaborative governance method, particularly relates to the field of audit data governance, is used for solving the problem of field blood relationship path fracture identification and repair, and comprises the following steps of: synchronously incorporating a running log and a metadata snapshot into a traceability system to form a stable field label mapping mechanism; field evolution logic is reconstructed in a structure diagram mode; the method comprises the following steps of: establishing a multi-source deviation correction coefficient of fusion structure and strength judgment by embedding two types of heterogeneous features of trend drift and chain break density, and guiding deviation path identification and weight regulation in a sub-graph isomorphism test, so that the evaluation of deviation risk does not depend on result comparison any more, but turns to the substantial analysis of an intermediate transformation chain; finally, potential field conflicts are actively recognized from distributed script evolution, a consanguinity consistency gap is automatically repaired, a credible chain type judgment basis is provided for cross-node auditing, and the interpretability and stability of the governance process are enhanced.

The invention discloses an electric forklift gantry synchronous inclination system and an inclination system control method, and relates to the technical field of vehicle engineering and electrical control, and the method comprises the following steps: establishing a time-frequency phase-locked observation baseline, and calibrating an abnormal trigger window by collecting a pressure wave change track of the gantry synchronous inclination system under a reverse pressure bearing condition; generating a pressure wave characteristic fingerprint and a time delay matrix as dynamic references for resonance path identification; and according to the pressure wave characteristic fingerprints and the time delay matrix, inverting a hydraulic pipeline reflection topology, identifying a pressure energy concentration area, extracting a resonance source anchor point, constructing a wave energy distribution diagram, and forming a space positioning basis of the standing wave risk. According to the method, hydraulic standing wave risk modeling and boundary extraction are achieved through time-frequency observation and resonance path recognition, a standing wave damping structure is constructed in combination with the phase conjugate test and track shaping technology, steady-state damping and high-precision synchronous inclination control of the portal under reverse pressure bearing are achieved, and the system stability and operation safety are improved.

The invention discloses a drainage basin heavy metal migration path identification method and system, and relates to the field of heavy metal detection, and the system comprises an unmanned aerial vehicle module which is used for sneaking in a drainage basin according to a preset path, and collecting the heavy metal content information in the sneaking process; the storage module is used for receiving the heavy metal content information collected by the operation of the unmanned aerial vehicle module in real time and storing the heavy metal content information; the branch decision tendency value is calculated based on multiple groups of near-node heavy metal data to select the sneaking direction, the migration path identification pertinence is improved, the risk is primarily judged through the maximum value of the single type of heavy metal during evaluation, then accurate judgment is performed through the characterization value, and the risk evaluation accuracy is ensured.

The invention discloses a welding seam path recognition and planning method and system based on lightweight Transform and multi-modal fusion. The method comprises the steps that point cloud preprocessing is conducted on collected three-dimensional point cloud data; explicit differential geometric structure information is extracted from a local neighborhood of the input point cloud and is used as an input or gating signal for enhancing the network learning ability; realizing saliency dynamic adjustment of a network perception focus along with a local geometric structure through a geometric gating attention mechanism, and outputting 3D features; based on a two-dimensional auxiliary sensing channel of BEV projection, BEV features of a weld path are extracted through a convolutional network; carrying out weighted fusion on the 3D features and BEV features, and carrying out path point regression; outputting a continuous and controllable three-dimensional track point sequence; and according to the three-dimensional track point sequence, weld path planning is executed. According to the method, the precision and the real-time performance are both considered, the structure self-adaption capability and the end-to-end deployability are achieved, and the method is suitable for various industrial welding scenes such as ships, steel structures and pipelines.

The invention relates to the technical field of state monitoring, in particular to a power transmission line fitting multi-element perception failure assessment method and system, and the method comprises the steps: obtaining a fitting response section, extracting starting and ending points, searching amplitude interruption, monitoring peak and valley changes, recognizing structure transition, calling front and back paths, recognizing an offset point, tracking direction fluctuation, extracting a main path, and extending a missing section. And comparing the trend access segment, the classification rhythm, the pairing jump and the repetitive response to obtain a fitting failure trend feature set. According to the method, the structure interruption position in the hardware fitting response curve is identified, the direction continuous section and the offset section are divided, the included angle sudden change point is extracted, the path extension relation is tracked, the jump area and the repeated response section are separated, the path loss caused by fluctuation fracture is complemented, and the interference of direction sudden change on trend judgment is weakened; and the turning position and the fluctuation rhythm are combined to identify the fatigue path trend, and the perception integrity and the path continuity of the failure trend under the multi-disturbance condition are enhanced.

The invention relates to the technical field of intelligent logistics and transportation management, and particularly discloses a multimodal transport goods source allocation system based on data analysis. The system comprises a data access and fusion module, a data quality governance and enhancement module, a time-space association and flow calculation module, a critical path and hub identification module and a dynamic transport capacity allocation decision module, and multi-source data integration, quality governance, time sequence association verification, critical path identification and dynamic transport capacity allocation are carried out. The accurate description of the flow direction of the goods source and the dynamic adaptation of the transport capacity resources are realized, and the operation efficiency and reliability of the multimodal transport system are improved.

This invention relates to a thermally constrained dynamic task scheduling method, apparatus, sensing and control device, and unmanned aerial vehicle (UAV), specifically within the field of electric vehicle sensing and control equipment technology. It aims to address issues such as inaccurate heat accumulation determination, weak task migration targeting, and easy timing conflicts in existing scheduling schemes. The method collects real-time equipment temperature data and sensing task processing volume, and determines the heat accumulation at the first processing end based on a safety temperature threshold range of 60℃-85℃. Based on core parameters such as the number of computational instructions and data throughput, it quantifies the task heat value through a calibration model and migrates lower-priority tasks to the second processing end according to their heat level. During task migration, the migration rate is dynamically adjusted, and the computation frequency of remaining tasks is fine-tuned according to the temperature decrease gradient. Task conflicts are avoided through timing-related boundary control. This invention achieves a balance between targeted cooling and task real-time performance, ensuring reliable execution of core tasks such as UAV path recognition and UAV speed detection.

This invention discloses a method, apparatus, system, and storage medium for relocating hospital departments, including: step S1, data modeling and parameter initialization; step S2, constructing a multi-objective optimization model; step S3, constructing a constraint system; step S4, dynamic priority calculation and critical path identification; step S5, relocation scheduling optimization solution; and step S6, generating the optimal relocation plan. The technical solution of this invention enables systematic optimization under multi-objective constraints, simultaneously considering economic cost, continuity of medical services, and resource utilization efficiency, while meeting the requirements of medical quality and safety.

The invention provides a network equipment firmware vulnerability authorization classification static analysis method and system, and the method comprises the steps: inputting an authorization identification function name and a return value definition into an authorization branch path basic block recognition module, and outputting a pre-authorization path root node basic block and a post-authorization path root node basic block, and inputting the path into an in-process unauthorized path identification module. Inputting an initial address of a to-be-classified basic block specified by a user and a binary program entry function into an inter-process unauthorized path identification module, and performing backward control flow analysis based on a function call relation graph; in the process, an in-process unauthorized path identification module is called to analyze whether each basic block to be analyzed in the function and a basic block of a root node of the function where the basic block is located have a pre-authorized path or not, and finally whether a pre-authorized path from the basic block to be classified to the root node of the binary program entry function exists or not is analyzed so as to determine the authorization attribute of the basic block to be classified. According to the method, the to-be-classified vulnerabilities can be authorized and classified, and the false alarm rate and the path explosion occurrence probability are reduced.

The invention discloses a robot path recognition method and system, and belongs to the technical field of robots, and the system comprises an acquisition module, a recognition module, a coffee cup shape recognition result acquisition module, a coffee cup position recognition result acquisition module, a determination module, a position posture acquisition module and a path recognition module. The method comprises the following steps: acquiring an environment image including a coffee making area, and performing multi-stage identification and processing on the image, including determining candidate block images of suspected coffee cups, obtaining coffee cup structure standard parameters for shape matching, and identifying a coffee making area range for position judgment. And finally, fusing the shape and position identification results to determine a target coffee cup so as to confirm the moving path of the coffee cup. The method can effectively solve the problem that in the prior art, due to the fact that coffee powder is attached to a camera lens, an image is blurred, coffee cup recognition is inaccurate, and path planning deviation is caused.

This invention relates to a train route identification method, apparatus, device, and medium, belonging to the field of rail transit technology. The method includes: acquiring train operation data and train network vector GIS data; inputting the train operation data and train network vector GIS data into a pre-trained graph neural network model, outputting feature data; associating and matching the feature data with a preset train network, and determining the train's corresponding route based on the latitude and longitude trajectory with the highest matching probability and a confidence level meeting a preset threshold. By using an improved graph neural network model to achieve deep fusion of train operation data and network GIS data, the association features between train operation and network topology are accurately captured, effectively solving the technical problem of mismatched train routes in complex network conditions.

The application relates to the technical field of welding automation and robot visual perception, and discloses a robot intelligent welding method and system, wherein the method comprises the following steps: constructing a three-dimensional point cloud image by adopting a laser triangulation method; filtering and enhancing the point cloud image, and extracting a weld point cloud set by adopting a U-Net semantic segmentation model; identifying a space path and regressing a pose based on the set to obtain a position-pose pair trajectory primitive; fitting the path by adopting a cubic B-spline curve and generating a continuous welding trajectory by interpolation; and combining visual feedback in the welding process, introducing proportional differential control logic to compensate for closed-loop trajectory errors. Compared with the technical problem that the existing technology cannot stably realize weld path identification and accurate tracking under complex structures such as curved surfaces and variable cross sections, the application realizes high-precision automatic welding of complex weld structures by constructing a perception control closed loop based on visual servoing, and improves trajectory fitting precision.