205 results about "A* search algorithm" patented technology

The invention provides a space field game behavior dynamic evolution and strategy deduction optimization method and system, and relates to the technical field of space field game behavior optimization, and the method comprises the steps: analyzing historical behavior data through a deep neural network, constructing a dynamic relation graph of participants, inputting a multi-agent reinforcement learning model, and outputting a game behavior model. Then, a multi-layer game scene evolution tree is constructed, the instant income and the long-term risk are calculated in combination with a dual neural network, dynamic updating is performed by using a search algorithm of a hierarchical attention mechanism, and candidate evolution paths are generated; and finally, evaluating the candidate paths by using a strategy evaluation model constructed by a space-time diagram neural network, selecting an optimal strategy path, and generating a strategy deployment scheme implemented by stages in combination with a knowledge inference engine. According to the method, the evolution trend of the game behavior in the space field can be effectively predicted, a feasible strategy deployment scheme is provided, and the scientificity and effectiveness of strategy decision are improved.

The invention discloses a document generation system and method based on multi-agent collaboration, and the method comprises the steps: understanding and extracting the high-level semantic information of a text input by a user, and dynamically generating a problem to guide the user to define a demand, so as to form a detailed document demand; recognizing the theme and key elements of the document from the detailed document requirements; the optimal sequence of task execution is determined based on a shortest path algorithm, potential conflicts in task execution are predicted and solved through parallel processing and a search algorithm, and optimized task planning is formed through feedback circulation; generating a document outline according to user requirements, retrieving associated contents of document chapters so as to write chapter contents, and performing grammar and format verification to form a preliminary document; defining a document quality evaluation index according to user requirements and industry standards, performing document quality evaluation on the preliminary document, and feeding back the document quality evaluation result to the user; and generating a complete final document and presenting the complete final document to the user. According to the invention, the document generation quality is improved.

Disclosed in the present application are an integrated circuit routing method and system, and a device and a medium. The method comprises: acquiring an integrated circuit layout model; performing routing area generation processing on the integrated circuit layout model on the basis of a particle swarm optimization algorithm, so as to obtain a routing area; performing routing processing on the routing area on the basis of an improved A* search algorithm, so as to obtain a routing path; on the basis of the routing path, removing a redundant segment, so as to obtain a post-processed path; and performing dictionary conversion processing on the post-processed path, so as to obtain a routing layout. In the embodiments of the present application, a discrete particle swarm optimization algorithm can be used to reduce a routing search space, and a detailed routing area is determined by means of an improved A* search algorithm, thereby further enhancing routing speed, and improving the routing efficiency of an integrated circuit. The embodiments of the present application can be widely applied to the field of integrated circuits.

The invention provides a quantum circuit optimization method, device and equipment for a ZUC algorithm subcomponent, and the method comprises the steps: encoding a signal transmitted by communication into quantum bits, and inputting the quantum bits into a ZUC algorithm structure for quantum implementation; representing a first linear subcomponent in the ZUC algorithm structure as a matrix in a binary field; after a matrix in the binary field is expressed as a block matrix, a low-dimensional matrix M in the binary field is derived from the block matrix, and quantum implementation of the M is searched by using a search algorithm designed based on a matrix decomposition principle; and according to a quantum gate parallel rule, carrying out optimized multi-layer division on the quantum of the M again, obtaining the quantum of the M, carrying out optimized layering operation, and optimizing the quantum of the first linear subcomponent. According to the method provided by the invention, the quantum circuit of the subcomponent of the ZUC algorithm is optimized, so that the resource overhead of quantum implementation of the ZUC algorithm is saved.

The invention discloses an intelligent vehicle automatic parking path planning optimization method, system and device and a storage medium. The method comprises the following steps: planning a parking path of a vehicle from a current position to a target parking space by using a search algorithm; a virtual parking environment is constructed, a simulation test is carried out in the virtual parking environment, different parking scenes and environment changes are simulated, and whether a parking path is safe or not is verified; when the parking path is verified to be safe, the vehicle is controlled to be parked according to the parking path; and when the vehicle is parked, the environment change is monitored in real time through a sensor, and the parking path is dynamically adjusted by using a path-finding algorithm and reinforcement learning. An A * algorithm is used for carrying out preliminary path planning and optimization, the reliability of the algorithm is verified through simulation testing, path planning and control strategies are continuously optimized, environment changes are monitored in real time and paths are dynamically adjusted by fusing data of various sensors such as a laser radar, a camera and an ultrasonic radar, and it is ensured that a vehicle safely avoids obstacles.

The invention relates to the field of electronic component workshop AGV logistics scheduling, and discloses an improved PPO scheduling method and system in an electronic component workshop AGV composite operation mode, a medium and equipment, and the method comprises the steps: carrying out the rasterization of a workshop environment map through grid modeling, enabling the AGV to carry out an A * search algorithm in a grid map through a plurality of task orders, obtaining batch sample data, and carrying out the calculation of the A * search algorithm in the grid map; pre-training is completed according to model parameters of the sample pre-training strategy network; according to the new task order, the AGV interacts with the environment, an observation state, an action space and a reward value are obtained, after matrix processing is carried out on the observation state, the observation state serves as input data of a PPO algorithm, a strategy network loss value and a value network loss value of the PPO algorithm are calculated, network model parameters are updated through a gradient descent method, and PPO training is completed; and generating a real-time task order on line, and guiding the AGV to complete path planning and scheduling tasks through the decision network models trained in the first two stages. According to the method, the electronic component workshop AGV scheduling strategy can be obtained through rapid learning, the learning efficiency is remarkably improved, and stability is achieved.

The invention discloses a method for deploying an edge dynamic DAG (Directed Acyclic Graph) server-free function on line to realize quick starting, and belongs to the field of edge computing. The method comprises the steps of 1, defining a dynamic DAG and establishing a modeling criterion of function preheating processing, and 2, constructing a dynamic DAG model with preheating preparation and function scheduling, and the model achieves the minimization of the overall expected execution time through preheating and function deployment. And step 3, converting a quadratic programming problem into a convex optimization problem by processing a quadratic constraint condition. And for the converted problem, a search algorithm based on random rounding is adopted to solve an optimal solution. And step 4, designing an online preheating and function scheduling algorithm. The algorithm includes function weight calculation, warm container adjustment, and optimal container selection to handle multiple online arrival requests. According to the method, through an online preheating and dynamic scheduling strategy, the cold start problem is effectively relieved, and the execution efficiency of the server-free function in the edge computing environment is improved.

A content channel generation device comprises a resource unit assignment circuit, for assigning scheduled station(s) as node(s) of a full binary tree according to a search algorithm; a node computing circuit, for determining first node connection information of the full binary tree, and to determine second node connection information of a smallest full binary tree according to a smallest binary tree algorithm and the first node connection information; a load balance circuit, for determining user field numbers corresponding to content channels according to a load balance function and the second node connection information; a user field generation circuit, for generating a traversal result of the smallest full binary tree according to a traversal algorithm and the second node connection information, and for generating user fields corresponding to the content channels according to the traversal result, to generate the content channels.

The invention discloses a quick video generation method and system based on structured space-time sparsity, and belongs to the field of computer vision and artificial intelligence. According to the method, an inherent structured and heterogeneous sparse mode of an attention matrix in a video diffusion model is systematically analyzed and revealed, and the core of the method is as follows: 1) a self-adaptive partitioning strategy is adopted, and various spatial interaction modes are approximated through dynamic image block grouping; 2) dynamically adjusting the sparse level of a window with time sequence change according to the inter-frame distance; and 3) automatically configuring a search algorithm, and optimizing a sparse mode while keeping a key attention path. The sparse attention mask optimized through offline pre-calculation is directly applied to attention calculation in a reasoning stage, so that 1.8-2.5 times of attention calculation acceleration is realized on a single GPU while the visual quality equivalent to a full attention baseline is kept, and a brand new solution path is provided for efficient long video generation.

The invention provides an AGV robot path planning method and system. The method comprises the steps that a map node graph, robot parameters, a starting node and a target node are acquired; introducing a robot constraint condition, and calculating an evaluation cost value of the robot from a current node to a target node based on a heuristic function; according to the evaluation cost value of the robot from the current node to the target node, based on an A * search algorithm, searching an optimal path of the robot from the starting node to the target node in map nodes; and generating a path instruction from the optimal path, and sending the path instruction to the robot for execution. Based on an advanced path planning algorithm, the path of the AGV robot is automatically planned, the optimal planning path is searched, and high efficiency of task execution of the AGV robot is achieved.

The invention discloses a cloud edge end heterogeneous resource cooperative training strategy based on hierarchical hybrid parallel, and aims to research how a cross-domain distributed computing framework fully utilizes various cloud edge end heterogeneous computing resources, research an equipment graph and computing graph construction mechanism, construct a simplified and merged computing graph and equipment graph, and reduce the search space of the hierarchical hybrid parallel strategy. A search algorithm of the hierarchical hybrid parallel algorithm is researched, a possible hierarchical hybrid parallel algorithm is searched through a dynamic programming algorithm and a resource constraint mechanism, and efficient hierarchical hybrid parallel in a cloud edge-end collaborative cross-domain scene is realized; a cross-domain hierarchical hybrid parallel optimization strategy during operation is studied, the calculation speed between nodes is adjusted through calculation of data loads in clusters and secondary balanced distribution of the calculation loads, and slow nodes during operation are avoided. By means of the mechanism, efficient adaptation of the cloud edge heterogeneous computing resources and the hybrid parallel strategy is achieved, the utilization efficiency of the cloud edge heterogeneous computing resources is improved, and the cross-domain distributed training speed of the model is increased.

The invention relates to the technical field of digital photoetching equipment, in particular to an LDI machine image segmentation and exposure splicing method, which comprises the following steps: carrying out primary segmentation on an original complex graph, and carrying out segmentation again after refinement processing; in the exposure process, periodically collecting a gray histogram of an exposure pattern, aligning the exposure images of adjacent areas to form a continuous exposure image, and dynamically adjusting exposure parameters by a light beam modulator according to the characteristics of each second segmentation area to realize closed-loop control on the exposure quality; all exposed images are spliced through a search algorithm and a multi-frame fusion technology, smooth transition of image edges is realized, discontinuity of local features is effectively avoided, the overall consistency and integrity of the spliced images are ensured, the imaging quality and precision are further improved, finally, the spliced images are verified, and the image quality is improved. And seamless splicing of each segmented region is ensured, so that high-quality graphic output is generated.

The invention relates to the technical field of laser spectrum gas detection, and particularly discloses a TDLAS (tunable diode laser absorption spectroscopy) double-gas concentration synchronous calibration and decoupling method in the presence of spectrum aliasing interference, and the method comprises the following steps: acquiring a multi-band harmonic signal subjected to adaptive optimization while acquiring environmental parameters; an optimal frequency band signal is determined based on the signal quality index; and calculating a spectral aliasing coefficient according to the characteristics of the optimal frequency band signal, dynamically adjusting operation parameters of a search algorithm based on a numerical value interval in which the spectral aliasing coefficient is located, and determining an initial range of dual-gas concentration in a search space. According to the method, the improved genetic algorithm based on the spectral aliasing coefficient dynamic adjustment search strategy is combined with the incremental support vector regression model, stage processing from coarse calibration to fine decoupling is achieved, and the calculation efficiency and the search precision can be automatically balanced when the spectral aliasing degree changes; and online correction is carried out on the model in combination with aging characteristics.

Systems, methods and interfaces are provided for processing natural language queries. A machine learning-powered search interface may process complex natural language queries across specialized domains. The method may utilize a transformer-based natural language processing model that analyzes user intent by examining both current and historical user inputs. A graph data structure may be used to represent interconnected domains like patient records, insurance claims, and business metrics enables dynamic information retrieval. The system may use an A* search algorithm to navigate domain nodes and identify relevant answers. An abbreviation expansion mechanism resolves technical shorthand by referencing domain-specific mapping databases, ensuring precise interpretation of user queries. The method may identify keywords, traces optimal answer pathways, and adapt by updating graph relationships and abbreviation mappings. By generating contextually rich natural language summaries, the system delivers precise, comprehensive responses through an intuitive search interface.

The invention relates to the technical field of unmanned aerial vehicle flight path planning, and discloses an unmanned aerial vehicle cluster four-dimensional flight path planning method for a high-density operation scene, and the method comprises the steps: constructing an urban airspace four-dimensional grid model, and completing the four-dimensional spatio-temporal data reconstruction of an unmanned aerial vehicle; the urban airspace four-dimensional grid model comprises a fixed four-dimensional body model, an unmanned aerial vehicle four-dimensional point model and a track four-dimensional line model; constructing a four-dimensional flight path planning model of the unmanned aerial vehicle; improving an adaptive heuristic function based on a 4D-A * search algorithm; and obtaining the optimal flight path of the unmanned aerial vehicle based on the operation cost and the operation risk obtained by the improved adaptive heuristic function. According to the method, high-density operation conflict-free real-time four-dimensional flight path planning of the urban unmanned aerial vehicle cluster can be effectively completed, the operation efficiency is improved, and the operation risk is reduced.

The invention relates to the technical field of automatic driving, and discloses a tramcar path planning method, device and equipment based on a reference line and a storage medium. According to the method, an improved A * search algorithm is combined with reference line discretization and smoothing processing, the node expansion number is reduced, the path smoothness is optimized, and the tramcar path planning efficiency is improved. Meanwhile, a double-factor heuristic function and dynamic constraint verification are utilized to ensure that the path is safe and feasible, and the method has the advantages of improving path planning efficiency, reducing path fluctuation, meeting special terrain and dynamic constraint requirements of a mining area and ensuring that the mine car stably travels along a predetermined reference line.

The invention relates to the technical field of service scheduling, and discloses a dynamic rule engine and state tracking service request automatic scheduling processing system. According to the system, scheduling is realized through a five-stage processing chain of request intake, state acquisition, rule fusion, decision generation and execution monitoring which are connected in sequence. The request intake stage captures a service request stream and generates a request feature topology; in the state acquisition stage, resource operation data is acquired in real time, and a state multi-dimensional map is constructed; in the rule fusion stage, time-space alignment is carried out on the two, and scheduling fitness is output through rule reasoning; in the decision generation stage, based on the scheduling fitness, a search algorithm is adopted to generate a scheduling command sequence; and tracking an execution result in the execution monitoring stage, calculating a scheduling deviation value and feeding back the scheduling deviation value to the request intake stage, and adjusting a feature topology generation parameter. According to the system, a closed-loop scheduling mechanism is formed, the dynamic adaptation of the service request and the resource state is realized, and the flexibility and pertinence of scheduling processing are improved.

The invention belongs to the technical field of fan arrangement, and discloses a fan point location automatic arrangement method and system based on artificial intelligence, and the method comprises the steps: firstly receiving a plurality of spatial constraint layers and deployment parameters, and unifying the coordinates; a conflict relation graph is constructed based on the processed constraint layers, nodes of the graph are different constraint layers, and edges are conflict strength among constraints; inputting the map into a pre-training map attention network, and reasoning to obtain a conflict slow-release factor which quantifies a constraint relaxation degree; then the factors are fed back to a point location search algorithm, and candidate point locations are evaluated by using a scoring function of the fusion factors; and finally, outputting a final arrangement scheme according to a scoring result. According to the method, through artificial intelligence middleware, such as a graph attention network, an artificial intelligence-based search algorithm and the like, the problem of excessive region deletion of a traditional method is solved, and the area and arrangement reasonability of a deployable region are improved.

The invention provides a Unity3D-based multi-agent collaborative integrated navigation virtual simulation system and method, belongs to the technical field of integrated navigation, and solves the problems that a traditional integrated navigation experiment method is high in cost and risk and is difficult to simulate various complex actual working conditions, and the simulation system comprises a display layer, a business layer and an underlying service layer; the display layer provides an operation and monitoring platform for a user; the business layer is used for performing control logic, queue management, path planning and artificial intelligence training functions on the agents in the simulation environment; and the underlying service is used for performing agent machine learning model training and project version management. The invention provides a low-cost and high-safety virtual simulation scheme for intelligent agent motion control and collaborative operation efficiency optimization, an improved A * search algorithm and an artificial intelligence training interface are integrated, and an economical and efficient virtual test platform is provided for research and development of an intelligent agent control algorithm.

The invention relates to the technical field of tunnel construction geological disaster early warning, and discloses a tunnel construction geological disaster early warning method and system. According to the method, multi-source geological monitoring data and environment dynamic parameters of a tunnel construction area are collected; and then through data fusion and feature engineering processing, data contradiction and redundancy are eliminated, and a standardized training data set is formed. And then training and verifying a disaster risk prediction model by using the data set and adopting a machine learning algorithm so as to obtain an optimized model to accurately predict the disaster risk. And performing risk value prediction on different geological structure units by applying the optimization model, generating a risk distribution map, and determining an optimal geological structure unit through a search algorithm. Based on this, numerical simulation of geological disaster triggering and propagation is executed, a disaster evolution rule is analyzed, and a path simulation data set is output. And finally, performing multi-objective optimization on the data set according to a safety early warning objective, generating a geological disaster early warning scheme, and performing iterative calibration.

The invention discloses an aircraft taxiing path automatic generation method and system, and the method comprises the steps: obtaining airport GIS map data, and constructing a taxiway hierarchy graph based on an H3 space grid; after an instruction containing a sliding path name sequence is received, the instruction is automatically integrated with machine position and target point information and mapped into starting and ending nodes in a graph structure; and furthermore, a search algorithm combining path name matching and turning angle punishment is executed in the graph, and a continuous and reasonable optimal geometric sliding path is automatically generated. According to the method, full-automatic mapping from a text instruction to a standardized geometric path is realized, the defects of dependence on manpower, low efficiency and high error rate in the prior art are overcome, and the intelligent level and operation safety of airport ground sliding guidance are remarkably improved.

Systems and processes for aligning weakly-annotated data to recognized characters in a document are provided. In a method for aligning annotation data to recognized characters, annotation words and character recognition tokens are received, and a search algorithm is performed to align the annotation words to the tokens in a stepwise manner. At each step, an annotation word is aligned to one or more tokens, and a cost of each respective alignment is calculated. Once all annotation words are aligned, a full set of annotation word-token pairs corresponding to the annotation is selected based on a total cost of alignment for that set. A bounding box enclosing the tokens in the selected full set is generated and output to a target application.

The invention provides a book recommendation method and system based on multi-agent reinforcement learning, and relates to the technical field of book recommendation, and the method comprises the steps: constructing a multi-agent reinforcement learning model, and achieving the information interaction and cooperation between agents through a centralized training architecture; based on the deterministic characteristics of the multi-agent reinforcement learning model, optimistic search is introduced to optimize book search; performing attention weighting optimization on strategy gradient updating through a dominant function to optimize a recommendation strategy; and the strategy parameters of the multiple agents are synchronized through the distributed execution architecture to realize collaborative decision making. According to the method, the multi-agent reinforcement learning models are constructed for mutual assistance and distributed execution, so that the calculation complexity is remarkably reduced, and real-time recommendation of large-scale users and books is supported. On the basis of the deterministic characteristics of the multi-agent reinforcement learning model, optimistic search is introduced to optimize book search, a search algorithm based on quantile optimistic estimation is proposed, and the search efficiency is remarkably improved. And the advantage value searched by the Monte Carlo tree is combined with the strategy gradient, so that the strategy degradation problem of the sparse reward scene is solved.

The invention discloses an intelligent adaptive personalized learning method and system based on AIGC and a knowledge graph. The technical problems that an existing online learning system is low in individuation degree, a knowledge system is fragmented, the evaluation dimension is single, the adaptive capacity is weak, and different requirements of teachers and students are not distinguished are solved. The method comprises the following steps: constructing and updating a domain knowledge graph library with weights, establishing a self-adaptive course resource library associated with knowledge entities, respectively generating multi-dimensional portraits of teachers and students, generating differentiated personalized learning / teaching paths based on an A * search algorithm, and realizing self-adaptive content recommendation, continuous self-adaptive evaluation and dynamic adjustment through a mixed recommendation algorithm. A closed-loop intelligent ecology is formed; the corresponding system comprises core modules such as a knowledge graph library, a self-adaptive course resource library and a teacher and student portrait module. According to the method, all-dimensional personalized support of teachers and students, dynamic self-adaptive adjustment and resource generation according to needs are realized, the teaching and learning efficiency is remarkably improved, and the method is suitable for practical subjects and has system self-evolution ability.

The invention discloses an evolved memory system and interaction method for a universal virtual agent in a GUI (Graphical User Interface) environment, and the system integrates sub-task nodes after decomposition of various tasks and the dependency relationship thereof by constructing a global memory graph, forms a memory graph, and generates a feasible GUI operation sequence by adopting a search algorithm based on the graph. And abstracting the sequence into a transferable guide rule by means of a large language model. When the task is executed, the matching rule and the subtask nodes are retrieved through semantic and structure mixing, the agent decision is assisted, the memory graph is updated after successful execution, and continuous evolution is achieved. When the system is practically applied, the intelligent agent can be driven to execute a GUI task through a natural language instruction, the intelligent agent generates an operation action by utilizing a rule and a retrieval mechanism in the memory graph, interacts with an interface to complete the task, and feeds back an interaction track which is successfully executed to the system, so that the structure of the memory graph is continuously optimized.

The invention discloses a method and system for cooperatively chasing a moving target by an unmanned surface vehicle cluster in an obstacle environment, and relates to the technical field of unmanned surface vehicles, and the method comprises the steps: obtaining the surrounding environment related information of the unmanned surface vehicles, carrying out the fusion processing of the surrounding environment related information of the unmanned surface vehicles, and obtaining the global environment information of a whole chasing region; based on the global environment information of the whole chasing area, adopting an algorithm to track and predict the position and the motion state of the moving target in real time to obtain the motion trail of the moving target in a future period of time, and based on the global environment information of the whole chasing area and the motion trail of the moving target in the future period of time, adopting a search algorithm to search the moving target in the future period of time; generating a chasing path of the unmanned ship; the tracking task is generated based on the tracking path of the unmanned ship, the related data of the unmanned ship is acquired, the tracking task is allocated based on the related data of the unmanned ship, the unmanned ship tracks the moving target based on the tracking task, and the tracking efficiency and success rate of the unmanned ship cluster in a complex environment can be improved.

The invention discloses a control method for a magic cube solving robot based on two arms and four fingers, and relates to the field of robot control. According to the method, the step optimization problem of mechanical action recovery of the robot with the two-arm and four-finger structure in the magic cube recovery task can be solved, and the robot can be allowed to perform the mechanical action recovery task on any magic cube in a disordered state. When a task is completed, firstly, a key operation surface and an operation stage are defined, then conversion logic from the magic cube restoration step to the mechanical execution step is constructed, and an optimal restoration path is explored by integrating reset and adjustment stages of adjacent steps and utilizing symmetric plane conversion and selection of an execution strategy. Thirdly, a pruning function is introduced, future step cost is predicted based on historical data, and search space complexity is reduced; and finally, the optimization function is integrated into a search algorithm, a reduction sequence with the least mechanical step number is searched, and the reduction sequence is accurately executed by a controller so as to efficiently and accurately complete the magic cube recovery task.

The invention discloses an urban motor vehicle diversified track generation method based on traffic road network weight learning. The method comprises the following steps: S1, generating a road network weight map of a traffic road network in a network layer; s2, based on the road network weight map, executing an A * search algorithm according to the starting point and the ending point of the vehicle, and outputting a vehicle reconstruction track; s3, reversely training a graph neural network GCN and a graph attention network GAT in a network layer according to the vehicle reconstruction trajectory and the loss of the corresponding real trajectory, and outputting a road network weight map with a connected vehicle trajectory; and S4, inputting the road network weight maps into a VAE model, obtaining a plurality of road network weight maps through coding, partition sampling and decoding processing and reconstruction in sequence, and executing an A * search algorithm on the road network weight maps to generate diversified vehicle tracks. The vehicle track generation method provided by the invention not only provides necessary data support and decision basis in the traffic management process, but also lays a solid foundation for future development of urban traffic.

The invention discloses a drainage basin nutrition load estimation method based on remote sensing data and machine learning. The method comprises the following steps: constructing target variables of a model, constructing characteristic variables of the model, selecting a combination of the characteristic variables by a search algorithm, establishing the model, and estimating an area to be measured by using the constructed model. According to the method, a GIS analysis tool, data acquisition on a GEE platform and a machine learning model are integrated together by using a python language. According to the method, feature variables are obtained through remote sensing data, drainage basin nutrition load estimation of a data-deficient area is achieved by building a machine learning model, and a scientific basis is provided for drainage basin management and non-point source pollution abatement.

The invention relates to the technical field of artificial intelligence, and particularly provides an indicator diagram automatic data enhancement method and device, and the method comprises the following steps: S1, alternately updating the parameters of a target model and an enhancement model based on an automatic enhancement method of teacher knowledge, the enhancement method comprises a color amplification model and a geometric amplification model, the target model is set as a teacher model, and the teacher model is any one model and only needs to be different from the target model; s2, setting a search algorithm and a search space, and updating the structure of the search space by using a color amplification model and a geometric amplification model; s3, generating a semantic auxiliary field; s4, performing image and semantic multi-modal fusion; and S5, preparing a data set. Compared with the prior art, the working condition diagnosis accuracy can be improved while the sample generation efficiency is improved, and the enterprise cost is reduced.