56results about How to "Avoid excessive dependence" patented technology

The invention relates to a multi-objective optimization scheduling method and system for a multi-AGV flexible job shop. The method comprises the steps of collecting job shop layout data, AGV operation parameters and production task data, and constructing a scheduling system basic model; performing recurrent neural network training and physical constraint model integration to obtain a hybrid scheduling prediction model; subtask decomposition and path search processing are carried out, and a dual-objective optimization model is obtained through constraint linearization conversion; performing particle swarm optimization solution on the dual-objective optimization model, and obtaining a non-dominated solution set through two-stage iterative calculation of an adaptive weight strategy and constraint conversion; performing multi-dimensional evaluation calculation on the non-dominated solution set to obtain an optimal scheduling execution scheme; and performing model feedback updating on the hybrid scheduling prediction model to obtain a target scheduling prediction model. According to the invention, multi-target balance of task completion time and energy consumption is realized, and the real-time response capability and operation stability of the system are improved.

The invention provides a landslide space prediction method based on a knowledge graph and representation learning. The landslide space prediction method specifically comprises the following steps: S1, constructing a landslide susceptibility knowledge graph considering mechanism knowledge and spatial-temporal characteristics; a knowledge graph ontology mode is designed, landslide disaster field knowledge modeling considering spatial-temporal characteristics and mechanism knowledge is completed, and on this basis, entities are extracted from multi-source data, relationships are constructed, and a knowledge graph data layer is constructed; and S2, through map semantic feature learning based on a ComplEx model, directed weighted graph construction based on a landslide susceptibility knowledge map, equivalent positive and negative landslide sampling based on a Louvain algorithm and map structural feature learning based on a GraphSAGE model, realizing landslide space prediction based on the knowledge map and representation learning. According to the method, diversified feature inductive learning under the constraint of fewer landslide samples in a complex heterogeneous environment can be realized, and the landslide space distribution probability is predicted through a node classification task.

The invention relates to the technical field of photovoltaic energy storage methods, and discloses a method for optimizing hybrid photovoltaic energy storage based on a Perturbobserve algorithm, which comprises a load, a photovoltaic power generation subsystem, an energy storage subsystem, a control and management system and a grid-connected system. The specific control method comprises the following steps: S1, obtaining an illumination intensity value within one day through a meteorological station, and obtaining a power grid net load value within one day through a voltage sensor and a current sensor; s2, taking economy optimization as a target, considering photovoltaic energy storage capacity constraints, and establishing an energy storage SOC control strategy target function; and S3, solving is carried out by using a PerturBobserve algorithm, and an optimal control pulse sequence is obtained. By optimizing the charging and discharging time and intensity of the energy storage battery, the consumption rate of photovoltaic power generation is maximized, the light abandoning phenomenon is reduced, and the energy loss in the charging and discharging process is reduced. Meanwhile, the PertuBobserve algorithm can avoid the excessive charging and discharging of the energy storage battery while considering the system load, prolongs the service life of the battery, and solves the problem of insufficient management of the service life of the battery in a traditional method.

The invention discloses a micro-service-based enterprise software system construction method and device, and relates to the technical field of enterprise software system construction, and the method comprises the steps: carrying out the micro-service division of an initial software system according to enterprise demand information and micro-service boundary information, obtaining an independent micro-service system, and carrying out the evaluation of the independent micro-service system, determining communication architecture information between the independent micro-service systems according to the micro-service configuration information, evaluating the independent micro-service systems to obtain a micro-service configuration influence index, acquiring the micro-service configuration information, evaluating the independent micro-service systems to obtain the micro-service configuration influence index, and determining the communication architecture information between the independent micro-service systems according to the micro-service configuration information. The micro-service architecture is evaluated and optimized through the micro-service configuration influence index and the micro-service independence index, and through quantitative evaluation and the influence index, the micro-service architecture can be judged and adjusted in time, the system design is optimized, and potential performance bottleneck and system excessive dependence are avoided.

The invention relates to the field of big data super-concurrent processing, in particular to an education big data super-concurrent processing system and method. The method comprises the following steps: collecting and preprocessing original request data to obtain actual request traffic; using historical data in actual request traffic to train a weighted autoregression model, and predicting the traffic; dynamically adjusting the concurrency based on the obtained flow prediction value, calculating the scheduling priority of the data packets, and sorting and distributing the data packets; a dynamic back pressure control mechanism based on flow prediction is introduced, flow input is dynamically adjusted, and overload is prevented. The problems that a traditional scheduling algorithm and a traditional load control strategy lack enough accurate flow prediction, concurrency cannot be adjusted in time, and system resources are excessively distributed or insufficient are solved; the priority processing of the key task cannot be ensured because the flow fluctuation and the scheduling priority are not fully considered; the problems that flexibility and self-adaptive capacity are lacked, the flow cannot be effectively limited when the load is too high, and the overload risk of the system is increased are solved.

The invention relates to the field of production scheduling and resource optimization, in particular to a flexible job shop scheduling method based on deep reinforcement learning, which is characterized by comprising the following steps: 1) constructing a flexible shop scheduling environment; (2) a flexible workshop scheduling environment is abstracted to be Disjunction Graph, and the flexible workshop scheduling environment is abstracted to be Disjunction Graph; 3) extracting state features; 4) coding the state features; 5) executing a decision based on the encoded features to obtain scheduling probability distribution; and 6) generating a scheduling plan according to the probability distribution. According to the invention, through multi-level feature extraction and fusion, accurate perception of a production environment is realized, and the accuracy and robustness of a scheduling decision are effectively improved; by utilizing a deep reinforcement learning framework, the dynamic change of the production environment can be adaptively learned, the scheduling strategy is continuously optimized, and the generalization ability is relatively high; and feature representation is enhanced through an attention mechanism, so that the model can still quickly find a global optimal solution when facing a large-scale and high-complexity FJSP problem.

The invention relates to the technical field of artificial intelligence, and discloses an AI-assisted data quality intelligent monitoring and analysis method. According to the AI-assisted data quality intelligent monitoring and analysis method, through a multi-step processing flow, multi-dimensional power data can be acquired in real time, timestamp marking can be performed, missing values and abnormal values can be effectively identified, missing value complementation is performed by adopting an interpolation method, so that data deviation is reduced, and in the feature extraction stage, the data quality is improved. The average power and the standard deviation of the power data are calculated, the volatility of the data is quantified, the data features are richer and more comprehensive, the power data quality prediction model generated through training of the support vector machine can comprehensively consider multiple dimensions such as the integrity, the accuracy, the consistency and the timeliness of the data for evaluation, and the evaluation efficiency is improved. Excessive dependence of a traditional method on a single index is avoided, and the real situation of power data quality is comprehensively reflected.

The invention discloses a zero-sample food image detection method based on visual semantic bidirectional guidance, which belongs to the technical field of image data processing and generation, takes DETR based on Transform as a detection framework, and fuses a food semantic enhancement network and a cross-modal bidirectional guidance module. Firstly, an image is input to an image backbone network branch to extract visual features, a category label is input to a food semantic enhancement network branch to extract semantic features, then the two obtained features are sent to a cross-modal two-way guide fusion module, the module comprises a two-way guide encoder, a two-way guide query vector and a two-way guide decoder, and the cross-modal two-way guide fusion module comprises a two-way guide query vector and a two-way guide decoder. Visual and semantic features are aligned and fused through coding and decoding processes, and finally, regression and classification tasks of zero sample food image detection are completed in a full connection layer. According to the method, detailed food attributes and visual features can be captured in the detection process, effective interaction of visual semantic modals is achieved, and therefore the detection precision of unseen food categories is improved.

The invention discloses an orthogonal DR-based adaptive planning method and related equipment, and relates to the technical field of radiotherapy, and the method comprises the steps: carrying out the automatic segmentation of a planned CT image of a patient, and obtaining a segmentation result; the segmentation result comprises a target region of the patient or a delineation set of an organ at risk; after a patient is positioned, DR images of the patient in two orthogonal directions are collected; projecting the planned CT image of the patient in the space of the positioning coordinate system to obtain a corresponding orthogonal DRR image; performing registration on the DR images in the two orthogonal directions and the corresponding DRR images to obtain 2D deformation grids in the two orthogonal directions; reconstructing a 3D deformation field through the 2D deformation grid; acting the 3D deformation field on the sketching set to obtain a deformation component of the corresponding target area or the endangered organ of the patient; determining whether to recollect the CT image of the patient according to the deformation component; the CT scanning frequency can be reduced, and meanwhile real-time evaluation of tumor changes is achieved.

The invention discloses an antibacterial magnetic compatible medical beta-type Zr-Nb-Cu alloy and a preparation method thereof. The alloy is prepared from, by mass, 18%-24% of Nb, 4%-10% of Cu and the balance Zr and inevitable impurity elements. The preparation method of the alloy comprises the following steps: 1, weighing sponge zirconium, copper and niobium particles as raw materials; 2, putting the raw materials into a water-cooled copper crucible, vacuumizing, introducing argon, and smelting in a non-self-consumption vacuum arc furnace; according to the beta-type Zr-Nb-Cu alloy, alloying elements Nb and Cu are selected, the strength, hardness, abrasion resistance and corrosion resistance of the alloy are improved while the elasticity modulus of the alloy is reduced, the lasting and efficient antibacterial performance of the alloy is guaranteed, the influence of postoperative bacterial infection and MRI detection artifacts is remarkably reduced, and therefore the beta-type Zr-Nb-Cu alloy has the lower elasticity modulus and magnetic susceptibility, and the service life of the alloy is prolonged. And the method is more suitable for human skeleton and nuclear magnetic resonance (MRI) detection.

The invention relates to a backup recovery method and device of a virtual machine, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: in response to any operation request for a backup set of a target virtual machine, starting a backup virtual machine in a host machine; the target virtual machine is a virtual machine to be backed up in the host machine; taking the virtual disk file in the backup set as a block device of the backup virtual machine, and loading the block device in the backup virtual machine; after the loading processing is completed, carrying out mounting processing on a virtual disk corresponding to the block device in a backup virtual machine, so that the backup virtual machine generates mounting information of the virtual disk; in response to a recovery request for the target file, recovering the target file by backing up the virtual machine and the mounting information; the target file is a file of the target virtual machine. By adopting the method, the backup recovery reliability of the virtual machine can be improved.

The invention discloses a high-precision low-cost seamless positioning system and method suitable for an intelligent sweeper. The invention provides a high-precision low-cost seamless positioning system and method suitable for an intelligent sweeper. The system comprises a three-dimensional laser radar, a GNSS, an IMU and a data processing unit. According to the method, dynamic and static states of a multi-source sensor are classified, GNSS and IMU data are combined in a dynamic state to perform pre-integration processing through extended Kalman filtering, positioning is performed by depending on a three-dimensional laser radar SLAM in a static state, and then the positioning precision is improved through an information fusion technology. And the data processing unit classifies and processes data according to the motion state of the sweeper and monitors and adjusts strategies in real time. The method effectively solves the problem of high-precision low-cost seamless positioning of the intelligent sweeper in different scenes, improves the sweeping efficiency and stability, and is suitable for the field of positioning and navigation of similar equipment such as the intelligent sweeper.

The invention discloses an asteroid structure estimation model based on spacecraft flight data and a prediction method thereof, and the method comprises the steps: directly predicting the internal structure of an asteroid through employing the position and speed sequence of a spacecraft, and combining a reversible neural network INNs, thereby constructing an asteroid structure estimation model, namely a DINNs model; the DINNs model is embedded in an on-board data processing center of the deep space exploration spacecraft, and when the DINNs model is applied to on-orbit real-time asteroid structure intelligent prediction, the spacecraft on-orbit asteroid structure prediction can be completed in a short time. According to the method, the asteroid structure can be predicted in real time, the detector can be helped to autonomously select high-value sampling points, and real-time decision making and autonomous control of the detector are served. The method provided by the invention has the characteristics of strong end-to-end fitting capability and relatively perfect framework and hardware foundation. According to the method, the end-to-end estimation of the parameters of the internal structure of the asteroid is realized by using DINNs for the first time, and a characterization extraction method for predicting the internal structure of the asteroid is provided.

This invention provides a method and system for UAV path planning based on a fusion algorithm, relating to the field of UAV autonomous navigation and path planning technology. The method includes the following steps: acquiring the starting coordinates, ending coordinates, and obstacle information in the flight environment of a flight mission, and constructing a flight environment spatial model; defining path constraints and optimization objectives, and generating an initial set of feasible paths based on the flight environment spatial model, path constraints, and optimization objectives; performing weighted fusion and guided enhancement on the initial set of feasible paths to obtain a weighted fused path and a guided enhanced path; employing a particle swarm optimization algorithm to construct an initial particle swarm based on the initial feasible path set, the weighted fused path, and the guided enhanced path, and iteratively optimizing the initial particle swarm to obtain a globally optimal particle; and decoding the globally optimal particle to obtain the final flight path. This invention significantly improves the efficiency and quality of UAV path planning in complex static obstacle environments.

The invention discloses a wireless network positioning method, server and system, and relates to the technical field of wireless communication and positioning navigation. The method comprises the following steps: receiving a signal sent by a to-be-positioned target node through a plurality of positioning reference nodes and extracting an original positioning parameter sequence; then, improved Z-score exception elimination based on median and median absolute deviation and adaptive attenuation coefficient smoothing based on average inter-frame hopping are carried out on the sequence, and a stable estimated value and a measurement variance are output; and finally, according to the number of the effective positioning reference nodes, adaptively switching a CHAN algorithm or a weighted least square method to carry out position calculation. The improved Z-score filtering and dynamic weight fusion technology is adopted, the defects in the prior art are overcome, the problem that a fixed algorithm is poor in robustness when the reference node topology change is positioned is solved through an algorithm self-adaptive switching mechanism, the precision and adaptability are remarkably improved in a complex scene, and meanwhile the deployment and maintenance cost is reduced.

The invention provides a software quality measurement method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a measurement demand index of a user for the quality of target software; determining a software measurement parameter for measuring the quality of the target software according to the measurement demand index; according to the software measurement parameter, determining a tool deployment relationship for measuring the target software, the tool deployment relationship being used for indicating an association relationship between a tool for realizing the software measurement parameter for measuring the target software and a measurement demand index; and capturing target data based on the tool deployment relationship and the software measurement parameters, and determining a software quality measurement result of the target software according to the target data. According to the method and the device, the measurement requirement of the target software can be butted with different software management tools, so that the software measurement is prevented from excessively depending on a certain tool, and the data acquisition flexibility and the data acquisition efficiency are improved.

The application relates to the field of artificial intelligence and computer vision, and provides a three-dimensional point cloud scene target positioning method based on a natural language instruction, which comprises the following steps: screening a relation triple containing an object category in a three-dimensional visual scene; determining a hierarchical attribute of the object category and constructing a hierarchical knowledge graph; analyzing a given natural language description to obtain a plurality of two-order semantic triples; adding the hierarchical attribute to the object in the two-order semantic triple to obtain a hierarchical two-order semantic triple; determining an initial visual feature of an object in a three-dimensional point cloud scene to be recognized; outputting a prediction score of the object; screening the objects corresponding to the prediction scores from high to low to obtain candidate objects; outputting a matching score of the candidate objects and a first high-order semantic triple; and determining a target object in the three-dimensional point cloud scene to be recognized. The application improves the prediction accuracy and has stronger robustness.

The invention belongs to the field of communication, and particularly relates to a centerless Mesh networking method, a communication node and a storage medium, and in the method, the communication node can join a corresponding network topology by sending a request of joining the network topology to any node of the network topology. When information needs to be transmitted to other nodes, the optimal transmission path can be automatically searched and determined for information transmission; in the application, the communication node can realize various node functions without being controlled by any central node, so that the excessive dependence of the communication node on a certain node is avoided, and the problem that each node in the network topology cannot work normally due to the fault of the certain node can be further avoided.

The application discloses a kind of industrial equipment's health degree analysis method without weight, belong to electric digital data processing technical field, comprising: in healthy state, according to the data index type for analyzing health degree, respectively collect at least one historical time window of industrial equipment industrial time series data, as reference data, collect the industrial time series data in current time window, as comparison data;According to reference data and comparison data, one by one through kernel density estimation, obtain probability density curve, to obtain probability distribution;According to data index type, the probability distribution of comparison data and the probability distribution of at least one reference data corresponding, obtain the JS divergence value for measuring data distribution similarity under data index type, to obtain the health degree of industrial equipment according to multiple JS divergence values.The application is based on kernel density estimation and JS divergence, quantifies distribution difference, without weight setting, improves health degree evaluation sensitivity and consistency.

The application belongs to the technical field of intelligent design and computing optimization, and provides a multi-agent-based cross-physical-field microstructure reverse design method and system, obtains a design intention, analyzes the design intention into a quantitative cross-physical-field attribute target vector; utilizes a pre-trained neural network model to map the design intention into an adjusting parameter in a latent representation space, decodes and generates a geometric configuration of an initial microstructure, and constructs an initial evolution population; performs cross-physical-field numerical simulation on the microstructure in the population; identifies an evolution direction of a parameter space through local gradient approximation, performs elite screening by using a non-dominated sorting driven by a Pareto, dynamically adjusts the adjusting parameter of the latent space, guides the evolution process to converge to a physically effective Pareto frontier, and obtains a design scheme according to an optimization result. The application realizes efficient and accurate design of a microstructure under complex constraints.

The present invention discloses an unsupervised microseismic denoising method and system based on a multi-channel Kalman filter network, which belongs to the intersection of earth science technology and artificial intelligence, including: obtaining multi-channel microseismic data and preprocessing; constructing a multi-channel Kalman seismic denoising network based on a recurrent neural network; constructing a loss function, and unsupervised training the multi-channel Kalman seismic denoising network based on the preprocessed multi-channel microseismic data and the loss function; denoising the real-time microseismic data through the trained multi-channel Kalman seismic denoising network. The present invention establishes a multi-channel Kalman filter state prediction and update equation, designs an RNN operation operator, realizes automatic parameter optimization, and improves the accuracy and efficiency of denoising. The method makes full use of the correlation of microseismic data, improves the filtering effect, is suitable for time-varying signals, and provides support for subsequent data processing.

The application discloses a clock error drift modeling method and system based on statistical likelihood optimization. The method first acquires clock error data in a clock synchronization system and performs standardization processing; then, clock speed and clock drift are calculated, and the clock drift is modeled as a random process conforming to a Gaussian process; next, a neural network model is constructed, the negative log-likelihood of the clock drift is taken as a loss function for model training, and the loss function is optimized to minimize the clock error drift error; finally, the system generates correction instructions by using the trained model, and high-precision clock synchronization is maintained when GNSS signals are missing. The application effectively improves the robustness and precision of clock synchronization, and is particularly suitable for environments where GNSS signals are unstable or unavailable.

This invention belongs to the field of sensors and signal processing, specifically relating to a wavelet kernel-scale sensitivity-guided denoising method for abrasive particle induced voltage signals. The method includes: acquiring abrasive particle induced voltage signals and performing harmonic cancellation to obtain a preprocessed signal; constructing a wavelet kernel function and calculating a kernel-scale guided spectrum using it and the preprocessed signal; constructing a sparse joint denoising model based on the kernel-scale guided spectrum; processing the sparse joint denoising model to obtain a convex optimization objective function; solving the convex optimization objective function using an adaptive step-size gradient descent method and an adaptive iterative shrinking threshold method to obtain a weight vector characterizing the distribution of abrasive particle characteristic signals; binarizing the weight vector characterizing the distribution of abrasive particle characteristic signals to obtain a feature indicator vector; performing a Hadamard product between the feature indicator vector and the preprocessed signal, followed by low-pass filtering to obtain a denoised signal. This invention can adaptively and non-destructively enhance and denoise abrasive particle characteristic signals under strong interference environments.

The present application relates to a pathological whole slice image classification method based on multi-branch independent mask and Dirichlet evidence fusion, belonging to the cross field of biological information and artificial intelligence. In view of the defects of traditional multi-instance learning method in weakly supervised classification task of pathological whole slice image, such as excessive attention concentration and static fusion, the present application sets dynamic mask parameters through multi-branch independent setting, forces different branches to pay attention to different pathological regions, and solves the problem of insufficient feature diversity caused by attention concentration; combining the confidence and uncertainty of Dirichlet distribution quantization branch prediction, the branch fusion weight is dynamically adjusted based on evidence theory, and the fusion robustness of multi-branch prediction result is improved. The experiment is verified on the public pathological data set such as CAMELYON-16, compared with the MIL method, the present application improves the AUC index by 1.1-2.4%, and significantly enhances the accuracy and generalization ability of pathological WSI classification.

This invention relates to the field of identity verification technology, specifically an identity verification method, system, device, and medium integrating multimodal biometrics. The method includes acquiring user features of a target user and extracting corresponding user feature vectors based on these features. The user features include facial biometrics, dynamic signature trajectory features, employee ID features, and password features. Based on the confidence level of each user feature vector, a target weight is determined for each user feature. The target feature vector is obtained by weighting each user feature vector and its corresponding target weight. Finally, the identity information of the target user is verified based on the target feature vector. The beneficial effects of this invention are that it not only effectively improves the system's adaptability and anti-counterfeiting level in complex application scenarios but also ensures high security and high availability of identity verification.

The invention discloses a preparation method of a self-supporting electrolytic water anode of a bimetallic site oxygen coupling electrolytic acidic medium oxygen evolution catalyst, and belongs to the technical field of electro-catalysis electrode preparation and electrochemistry. The method comprises the following steps: firstly preparing a 2.0 M nitrate aqueous solution, mixing according to a non-noble metal equimolar ratio, taking the mixed solution, adding the mixed solution into a PE (Poly Ethylene) ball milling tank for ball milling, mixing the ball-milled solution with ethanol for ultrasonic treatment to obtain a catalytic solution, smearing the catalytic solution on a carrier, drying under an infrared lamp, and finally annealing in a muffle furnace. According to the method, precious metal is not needed, the prepared catalyst self-supporting electrolytic water anode is excellent in OER performance, durability, energy conversion efficiency and the like in the acid environment, the preparation process is simple, raw materials are easy to obtain, the product is stable, and the industrial application potential is achieved.

This invention discloses a dual-lifter coordinated conveying control method applied to a conveying system. The conveying system includes a first conveying section, a first lift, a second conveying section, a second lift, and a control system. The method includes: before a material carrier enters the first lift from the first conveying section, the control system controls the first lift to send a reservation request to the second lift; if the second lift meets preset authorization conditions, it provides authorization permission and locks the entrance of the second lift to prevent the material carrier from entering; after obtaining authorization permission, the material carrier enters the first lift from the first conveying section, and the first and second lifts dock at the same height; the conveying mechanisms of the first and second lifts are synchronously driven to transfer the material carrier from the first lift to the second lift; after the material carrier has completely entered the second lift, the lock on the entrance of the second lift is released, and the material carrier is sent from the second lift to the second conveying section.

The application discloses a kind of uncertainty field environment under multi-space-ground robot cooperative search method, it is related to multi-agent cooperative search technical field.The method includes: grid field task area, define space-ground robot state;Based on large language model, initialize the search value of task area cell;Design the total cost of cooperative search path planning including path cost and mode switching cost;Introduce communication distance constraint, adopt path decision method based on predictive control, select the minimum path point motion of cost, and dynamically update search value to carry out rolling optimization.The present application reduces invalid search area using large language model, improves search efficiency;Mode switching cost and different mode energy consumption are included in planning, avoid frequent switching and excessive dependence on high energy consumption mode, improve endurance.The present application is suitable for heterogeneous or homogeneous multi-agent system, significantly improves the practicability and efficiency of multi-space-ground robot cooperative search.

The invention discloses a motor electric control anti-galloping protection device based on an electric tricycle, which comprises a support, a motor, a transmission shaft, a driving mechanism and a brake mechanism, the brake mechanism is connected with the transmission shaft, the driving mechanism comprises a driving shaft, a main gear, an auxiliary gear and an extension assembly, one side of the main gear is fixedly connected to the other end of the driving shaft, and the other side of the main gear is fixedly connected to the extension assembly. The auxiliary gear fixedly sleeves the outer wall of the transmission shaft, the main gear is meshed with the auxiliary gear, and the extension assembly is mounted on the other side of the main gear. Therefore, excessive dependence of a traditional technology on a precision sensor is avoided, potential safety hazards caused by sensor faults or collision damage are effectively reduced, and the reliable protection effect can still be achieved under more extreme operation conditions. Meanwhile, the structure of the main gear, the auxiliary gear and the supporting spring which are high in abrasion resistance is ingeniously integrated in the device design, durability and stability under frequent use are guaranteed, and therefore the device adapts to the long-time and high-strength tricycle use scene.

The invention provides a special-shaped household garbage classification method and system based on a knowledge graph, and relates to the technical field of artificial intelligence, and the method comprises the steps: collecting multi-dimensional parameter data of special-shaped household garbage, executing preprocessing and feature extraction to generate visual and material features, integrating and generating structured feature vectors, and converting the structured feature vectors into semantic feature representations, semantic feature representation is generated; constructing a junk knowledge graph, performing feature matching on the semantic feature representation and the knowledge graph, calculating an entity comprehensive matching degree, and screening to generate an instance candidate entity set; based on the instance candidate entity set and the knowledge graph, analyzing and reasoning an association path of the candidate entity and the garbage category, and calculating a credibility score of the association path; according to the method, the entity comprehensive matching degree, the association path and the credibility score are evaluated, and the garbage category is output, so that accurate classification of the special-shaped household garbage can be realized, the limitation of traditional matching based on a shape template is broken through, and the accuracy and adaptability of garbage classification are improved.