935 results about "Data preparation" patented technology

The invention discloses a building elevator detection, diagnosis and decision-making method based on a graph retrieval enhanced agent. The method comprises the steps that 1, elevator detection data are prepared and processed; step 2), knowledge extraction; step 3), knowledge fusion; step 4), visualization and optimization of the knowledge graph; 5) performing graph retrieval enhancement generation; step 6), diagnosing a decision-making agent; according to the method, triple information can be extracted from structural data, text data, visual data and other multi-modal data in the elevator detection field by guiding a multi-modal large model through an elevator detection technical specification, and an elevator detection visual target entity and a text named entity are automatically aligned based on a pre-trained vision-language model; the multi-modal knowledge graph in the field of elevator detection is accurately and efficiently generated, and building elevator detection intelligent diagnosis is carried out on the basis of the multi-modal knowledge graph and the fusion graph retrieval enhancement technology.

A method based on a multi-task learning model comprises the following steps: acquiring and processing histopathological image data of gastritis through a data preparation and preprocessing step; feature extraction is performed by using a self-supervised learning pre-trained model, and image blocks are coded into high-dimensional feature vectors; and constructing a multi-task learning model, learning feature representation through a full connection layer module and an attention layer module, and outputting a classification result of each task. And carrying out model training and optimization by using an optimizer, adding multitask loss through a loss function, and dynamically adjusting the model performance. The trained model can automatically detect and grade gastritis, atrophy, acute activity, intestinal metaplasia and other pathological indexes, outputs a standardized evaluation result, and assists in pathological diagnosis. According to the method, a multi-task deep learning framework based on self-supervised learning pre-training is constructed, a traditional single-task modeling mode is broken through, deep learning framework design is driven through pathological index association, and accuracy and clinical practicability of non-neoplastic disease assessment of the stomach are remarkably improved.

The invention discloses a water quality time sequence prediction method of an SSA-VMD-LSTM-XGBoost hybrid model, and belongs to the technical field of water quality monitoring and prediction. Comprising the following steps: (1) data preparation and preprocessing; (2) optimizing the water quality time sequence decomposition of the VMD based on SSA: optimizing a penalty factor and a modal number of the VMD by adopting a sparrow search algorithm (SSA), and decomposing the water quality time sequence into a plurality of sub-components with high stability and low complexity by utilizing the optimized VMD; (3) construction and training of an LSTM-XGBoost hybrid prediction model: constructing a hybrid prediction model fusing long-short term memory (LSTM) and extreme gradient boost (XGBoost), inputting a high-frequency component into the LSTM model, inputting a low-frequency component into the XGBoost model, and finally performing superposition and integration on prediction results of the models; and (4) multi-component prediction result integration and performance verification. According to the method, adaptive optimization of VMD parameters is realized through SSA, the feature extraction and time sequence modeling capability is improved by combining the advantages of LSTM and XGBoost, and the prediction precision and stability of the water quality time sequence are effectively improved.

The invention relates to the technical field of traffic safety, in particular to a beforehand prevention and control strategy generation method for traffic safety risks of a highway network in a mountainous area. Comprising the following steps: risk diagnosis and data preparation: collecting road basic information, historical traffic flow data, meteorological data and accident-prone point data of a mountainous area expressway network; constructing a risk assessment model: selecting road alignment, traffic flow, weather and environmental toughness indexes, determining index weights by adopting an improved analytic hierarchy process, constructing the risk assessment model through a fuzzy comprehensive evaluation method, and calculating the traffic safety risk level of each road section; a prevention and control strategy is generated; and strategy verification and iteration. According to the method, the total factor evaluation model is constructed by integrating the road network topological structure, the traffic flow characteristics and the meteorological sensitive section data, so that the problem of one-sided risk identification caused by single factor analysis in the prior art is solved, and systematic description of the mountainous area highway composite risk scene is realized.

The invention relates to the technical field of laser surface modification, and discloses a water turbine laser cladding process parameter intelligent optimization system and method, and the system comprises a data preparation and preprocessing module which is used for obtaining and preprocessing water turbine laser cladding process parameters and corresponding cladding layer quality index data; a physical information guided multi-task prediction network construction and training module; a causal inference analysis module; and a causal perception multi-objective optimization algorithm module. By constructing a multi-task prediction network guided by physical information, the internal physical law of the laser cladding process is fused into a deep learning model, and the accuracy and reliability of predicting a plurality of key quality indexes of the laser cladding layer of the water turbine are remarkably improved. Compared with a traditional pure data driving model, the method can generate a prediction result more conforming to an actual cladding mechanism by means of guidance of physical constraints even under the condition of limited experimental data, and lays a solid data foundation for subsequent process optimization.

The invention provides an SQL statement analysis method and device, equipment and a storage medium, and relates to the technical field of databases. According to the specific scheme, the method comprises the steps that SQL statements and metadata thereof are obtained, wherein the metadata comprises table names, column names and predicates of the SQL statements; target statistical information corresponding to the metadata is obtained from a preset statistical information base, and the target statistical information comprises table statistical information, column statistical information and index statistical information. And loading the statistical information to a test environment database to obtain a simulation database corresponding to the SQL statement. And based on the statistical information, determining an execution plan of the SQL statement, and executing the SQL statement in the simulation database according to the execution plan to obtain a statement analysis result. The method can effectively solve the problem of analysis distortion caused by the distribution difference of test and production environment data, and can obtain a more accurate SQL analysis result without complex data preparation.

The invention relates to the technical field of data preparation pipeline recommendation, in particular to a context semantic perception-oriented data preparation pipeline recommendation method and system. The method comprises the following steps: performing data preprocessing on an obtained automatic pipeline construction data set; aiming at the preprocessed data set, carrying out double-view feature fusion based on a self-attention mechanism; for the preprocessed data set, context semantic information extraction based on a large language model is carried out; based on feature fusion and information extraction, a pipeline recommendation result is obtained through a deep reinforcement learning network; evaluating and optimizing a pipeline recommendation result; and an output pipeline. Through double-view feature fusion based on a self-attention mechanism, end-to-end dynamic integration is carried out on statistical attributes of a data set and a historical pipeline component sequence.

The present application relates to the technical field of electric digital data processing, and provides a data preparation method, system, and device for AIGC-based interaction analysis, and a medium. The method comprises: by means of a pre-trained text label recognition model, determining a label sequence corresponding to an interaction statement having experienced first word segmentation processing; on the basis of a preset expression construction rule, determining a label combination sequence corresponding to the label sequence and a conditional expression corresponding to the label combination sequence; by means of a preconfigured database candidate text dictionary and a preset matching rule, determining a strictly matched keyword and a fuzzily matched keyword corresponding to the interaction statement having experienced second word segmentation processing; on the basis of the strictly matched keyword and a similar candidate text corresponding to the fuzzily matched keyword, determining a word segmentation result corresponding to the second word segmentation processing; on the basis of a plurality of preset SQL statement rule matching templates, generating an interaction SQL query statement, so as to determine, on the basis of the interaction SQL query statement, interaction response information corresponding to the interaction statement.

The invention discloses a drug-drug interaction prediction method based on a drug flow sub-graph, and relates to the technical field of bioinformatics, and the method comprises the steps: data preparation: collecting a reference data set including drug-drug interaction, and introducing an external knowledge graph for adaptability preprocessing; constructing a model: constructing a drug-drug interaction prediction model, and training by using the preprocessed reference data set; and effect prediction: inputting a target drug into the drug-drug interaction prediction model, and outputting a drug-drug interaction prediction result through the drug-drug interaction prediction model. The structure and semantic information of the drug flow sub-graph are fully utilized, accurate prediction of drug interaction is realized, and the prediction efficiency is improved. And the interpretability of the drug-drug interaction prediction model is improved.

The invention provides a data weaving platform-oriented self-service data preparation method, which comprises the following steps of: accessing original data and associated structured metadata of a target data source into a data weaving platform to generate standardized accessed original data and associated structured metadata; based on a semantic model of the data weaving platform, performing classified semantic annotation on the associated structured metadata, and positioning a transfer node of the associated structured metadata to generate a data blood relationship record; performing bidirectional quality verification on the data consanguinity record according to semantics and path continuity, and matching with a federated governance rule of a data weaving platform to generate a metadata governance list; and generating a unified metadata directory based on the metadata governance list, so that a user of the data knitting platform can obtain key segments containing complete data consanguinity records. According to the method, the problem of insufficient accuracy and efficiency in self-service data preparation can be effectively relieved, and the efficient preparation requirement of a data weaving platform for multi-source data is met.

Disclosed is an explainable method for monitoring a state of a generator of a wind turbine generator system on the basis of a spatio-temporal graph. The method includes: S1: acquiring data collected by a supervisory control and data acquisition (SCADA) system; S2: carrying out data understanding on the SCADA data, selecting features associated with the generator, and carrying out data preparation on the selected feature data, and obtaining valid data; S3: embedding the SCADA data, and forming a directed spatio-temporal graph data sequence; and S4: carrying out modeling of a normal behavior model of the generator on the constructed directed spatio-temporal graph data sequence, computing a full-graph-level residual and a node-level residual, computing a residual through an exponentially weighted moving average (EWMA) control chart method, carrying out full-graph-level state monitoring on the generator, forming a fault information transmission chain relation, and enhancing explainability and robustness of a monitoring result.

The invention provides a counterfeited audio detection system and method based on mask reconstruction and time-frequency feature fusion, and the system comprises a data preparation module, a spectrum feature extraction module, a time-frequency feature fusion and prediction module, a training module, and a counterfeited audio detection module. Through a double-branch structure of an audio mask auto-encoder, features are extracted from two angles of reconstruction learning and direct encoding, enhancement and sequence correlation modeling are performed on the double-branch features, the ability of capturing forged clues is enhanced, frequency domain information and time domain context are combined, fusion features with higher discriminative ability are formed, and the accuracy and the reliability of the audio mask auto-encoder are improved. Through unified optimization of the total loss function, effective learning extraction and fusion of audio features are enhanced, the accuracy of forged audio detection is remarkably improved, and the method is suitable for actual forged audio detection application.

The invention relates to a construction organization design word document intelligent generation method and system. The method comprises the steps of establishing a preset template, loading and displaying the preset template, loading data, replacing placeholders, generating a temporary intermediate scheme file, editing communication settings in a web mode, editing the temporary intermediate scheme file, editing cover information of a scheme file to be generated, and generating and storing a final document. According to the method, the problems of poor template reusability, disordered format, low collaboration efficiency, low compiling efficiency, version conflict and the like in traditional construction scheme compiling are solved, intelligent and standardized generation of construction organization design documents is realized, multi-person collaboration data editing operation can be realized through editing of Excel structured data, and the method is high in practicability and easy to popularize. The data preparation difficulty of scheme editors is greatly reduced, the team efficiency is improved, and the data editing efficiency is also improved through multi-person cooperation.

The invention particularly relates to a method and a system for automatically generating legal instruments through a large model. The method comprises the following steps: preparing and marking data; training a large model; information processing; generating a document; and outputting and feeding back. According to the method, large model training and a multi-dimensional data processing mode are combined, so that automatic generation of the legal document is realized; in the data preparation stage, comprehensiveness and accuracy of legal knowledge are ensured through hierarchical labeling and knowledge base construction; a plurality of evaluation indexes, optimization algorithms and regularization means are applied in large model training, so that the model can accurately learn a legal language mode and a logic relationship; in the information processing link, the case information input by the user can be deeply understood by means of the natural language processing technology and knowledge graph construction, the legal document writing time is greatly shortened through the series of operations, errors and omissions caused by manual writing are reduced, and the legal document generation efficiency and quality are remarkably improved.

Disclosed in the present application are an aircraft final assembly and integration testing platform and testing method. On the basis of a rational platform architecture, structured process data preparation, model-based operation guideline document preparation, and dispatch assistance decision-making are realized by means of an application layer, such that uniform data sources in the field of final assembly can be achieved. Data uploading and issuing between an industrial control network and an internal network, and data interaction control between layers are implemented by means of the mutually collaborative operation of the application layer, a storage layer, an edge layer and an execution layer, thereby improving the information acquisition efficiency and the operating efficiency. The edge layer and the execution layer collaboratively operate, so as to realize the interactive collaboration between the assembly guidance work for operators and automatic testing, thereby improving the efficiency of data information acquisition, operation instruction and fault handling during a final assembly and testing process of an aircraft.

The invention discloses a remote sensing image change detection method based on deep learning and a GIS, and relates to the technical field of image processing and geographic information system application. Comprising the steps of 1, carrying out data preparation and preprocessing, 2, constructing a deep learning network model, carrying out feature extraction on preprocessed remote sensing image data by adopting a multi-layer convolution layer and a pooling layer, and aiming at obtained GIS data, carrying out feature extraction on the remote sensing image data by adopting different feature extraction methods according to data types, 3, performing feature fusion: fusing the remote sensing image feature vector and the GIS feature vector in a deep learning network model by using a feature fusion layer to obtain a fused feature vector; and step 5, performing change detection and result output: applying the trained deep learning network model to the preprocessed to-be-detected remote sensing image data and GIS data, outputting a change probability value of each pixel point through forward propagation calculation of the model to obtain a change probability graph, and performing post-processing operation on an obtained binary change detection result to obtain a change probability graph. And finally, outputting a clear and accurate remote sensing image data change detection result, and displaying the position, range and type information of a change area in the form of image or vector data.

The invention belongs to the technical field of intelligent manufacturing of aeronautical manufacturing enterprises, and relates to an assembly process procedure auxiliary compilation method based on a large model, which comprises the following steps: data preparation and processing: collecting and preprocessing an assembly process procedure text file; carrying out algorithm model selection and fine tuning, using an assembly process procedure text as a training data set and a test data set, and selecting different types of large models to carry out fine tuning; service publishing: deploying a plurality of large models with the best evaluation indexes as services, publishing the services, and externally providing available interface services; constructing an RAG knowledge base, vectorizing the assembly process procedure text, and constructing a vector knowledge base; intelligent question and answer: performing knowledge question and answer based on the RAG knowledge base; and assembly process procedure file generation: constructing the assembly process procedure file through text generation. According to the method, the functions of semantic retrieval intelligent question answering of the assembly technological procedure knowledge base, automatic assembly technological procedure file generation, information recording and the like are achieved.

The invention provides a method and device for an intelligent computing center cloud platform to adjust model training parameters according to computing power operation states, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructures. The first working component is a component deployed in a node in the intelligent computing center cloud platform, the target link comprises a data preparation link, a model training link, a model evaluation link and / or a model test link in the first working component, and the target link performs data processing or model training based on preset hyper-parameters; s2, calculating target correction parameters corresponding to the operation related parameters in a preset period based on a preset regulation and control model; and S3, adjusting the preset hyper-parameter based on the target correction parameter to obtain a target hyper-parameter. The computing power utilization rate of the intelligent computing center cloud platform can be greatly improved, and then the model training efficiency is greatly improved.

The invention discloses an intelligent bone marrow megakaryocyte morphology detection method which comprises the following steps: S1, data preparation: collecting and preprocessing a digital large map of a bone marrow smear, labeling megakaryocytes, and establishing a labeled sample; s2, generating and sorting a sample: extracting a small graph sample by taking megakaryocyte as a center; s3, constructing a deep learning model: constructing and training a deep convolutional neural network, and performing model optimization and performance improvement by using the generated small image sample set; s4, large image detection and reasoning: calling the trained model for reasoning by adopting a sliding window mechanism, and fusing detection results of a plurality of small windows back to an original large image through confidence weighting and a non-maximum suppression strategy; and S5, target cell segmentation: carrying out target segmentation on the detected megakaryocyte, and introducing a pyramid structure for cells with different sizes to obtain an accurate segmentation mask of each target cell. According to the method, the megakaryocyte detection and segmentation precision is improved through large image labeling, small image training and a sliding window reasoning strategy.

The invention provides a multi-LoRA large language model deployment system based on a cloud computing platform. A cloud computing AI platform layer trains a LoRA adapter matched with a basic large language model for a reasoning process; the multi-LoRA dynamic loading layer dynamically switches LoRA adapters needing to be mounted according to the request parameters, GPU optimization configuration is carried out according to service priorities in the request parameters, a basic large language model is loaded, and a plurality of LoRA adapters stacked in a sparse matrix form are mounted; and the resource scheduling optimization layer responds to the resource scheduling application, and outputs the request parameters subjected to priority management and hardware sensing optimization to the LoRA dynamic loading layer based on the container arrangement platform. According to the method, LoRA and container arrangement are deeply fused, and an automatic assembly line of training and reasoning is achieved. A video memory sharing mechanism enables a plurality of service scenes to share the same basic large language model, a LoRA adapter is loaded as required, and video memory occupation is greatly reduced. Full-life-cycle management from data preparation to model service is supported, and the large model deployment cost is remarkably reduced.

The invention relates to a multi-class archive knowledge question and answer agent based on a large language model, and belongs to the technical field of artificial intelligence and archive information management. The intelligent agent comprises a data preparation module, a semantic knowledge base construction module, a perception and analysis module, a decision and matching module, an execution and retrieval module, a cognition and reasoning module and an interaction and generation module. According to the method, the complex query intention of the user can be accurately understood, automatic association and semantic reasoning across archive categories are achieved, and the semantic splitting problem of a traditional retrieval mode in cross-category and multi-level query is effectively solved.

The invention discloses an intelligent data analysis workflow system and method based on ChatBI, and relates to the technical field of natural language processing and intelligent data analysis. A workflow stage is divided, and system cue words are provided according to the workflow stage, and the method comprises the following steps: an analysis intention recognition stage, establishment of different semantic models, arrangement of the semantic models through the system cue words, generation of initial analysis dimensions in combination with historical problems or agents, an intention clarification stage, and demand adjustment according to user feedback. The method comprises the following steps: verifying the rationality of operation through an intelligent agent and generating an adjustment suggestion, which comprises the steps of newly adding or merging indexes, performing a data preparation stage, executing data type cleaning, performing dimension and index correction, optimizing a data set by the intelligent agent according to domain knowledge, performing a report production stage, generating a chart configuration template, and recommending a chart type and an analysis model by the intelligent agent. A large model selects an analysis model, a preliminary analysis summary is generated, a report optimization stage is carried out, a user puts forward modification suggestions and intelligent agent verification logic, a chart configuration template is regenerated, and analysis content is combined.

The invention relates to the technical field of disease diagnosis, in particular to a sacroiliac joint lesion intelligent identification method and system based on MRI, and the method comprises the steps: data preparation and enhancement, multi-expert collaborative labeling, mixed 3D-2D model processing, verification and optimization, and deployment and interaction. Compared with the prior art that a scheme for detecting the sacroiliac joint lesion by adopting a single 3D CNN or 2D CNN model has the problems of inaccurate positioning of an anatomical structure and loss of inter-layer feature association, resulting in insufficient sensitivity and specificity, the method provided by the invention has the advantages that through an improved hybrid 3D-2D model architecture, spatial features are extracted by utilizing the 3D CNN and an inter-layer dependency relationship is modeled in combination with Transform, so that the sensitivity and the specificity of the sacroiliac joint lesion detection are improved, and the detection accuracy of the sacroiliac joint lesion detection is improved. Meanwhile, spatial pyramid pooling and a channel attention module are adopted to enhance the multi-scale feature extraction capability, so that the method has the advantages that the lesion detection sensitivity and specificity are remarkably improved, the bone marrow edema region can be more accurately recognized, and the misdiagnosis rate is reduced.

The invention provides a reservoir area landslide susceptibility evaluation method considering reservoir water level change factors. The method comprises the following steps: preparing data; constructing a model; introducing an adjustment coefficient; probability calculation and loss function definition; optimizing parameters; carrying out regularization processing; and outputting a result. The invention mainly considers the influence of reservoir water level change on reservoir area landslide susceptibility evaluation, provides a reservoir area landslide dynamic susceptibility evaluation method considering reservoir water level change, and mainly solves the following problems: a traditional landslide susceptibility evaluation technology is mainly oriented to administrative division, and the method mainly aims at evaluating the reservoir area landslide susceptibility, so that the reservoir area landslide dynamic susceptibility evaluation method can be used for evaluating the reservoir area landslide susceptibility. A river distance is introduced as an important influence factor of reservoir area landslide susceptibility evaluation.

The invention discloses a method for applying linear programming to CDN (Content Delivery Network) scheduling, which relates to the technical field of content delivery networks and comprises the steps of data preparation, strategy layer version smooth configuration, macroscopic layer and microscopic layer linear solution and online execution. Basic data are collected, cleaned and repaired, and a version change rule is set; the macroscopic layer constructs a linear programming model, and the cross-provincial bearing quota is solved with the aim of minimizing the cross-provincial cost; the micro layer takes the quota as a boundary and generates domain name class-node weight vectors in parallel; and adapting a routing request online through weighted rendezvous hashing and request features. According to the method, a dynamic cost matrix and a weight granularity control technology are integrated, the engineering problem of linear programming is solved, second-level response, approximate global optimal scheduling and accurate execution of floating-point-level weight are realized, memory overhead is reduced, smooth updating of a strategy and system stability are guaranteed, and CDN service quality and operation efficiency are improved.

The invention belongs to the technical field of computer vision, and discloses a target identification tracking method based on a self-supervision mechanism, and the method comprises the steps: enhancing a self-supervision pre-training module through causality, constructing a causal sample pair through unlabeled video data, learning universal features through combining with comparison loss, and achieving the high-precision tracking without large-scale manual labeling. A multi-modal feature fusion and dynamic calibration mechanism further reduces dependence on annotated data, is especially suitable for industrial inspection, field monitoring and other scenes where data acquisition is difficult, significantly reduces time and labor costs in a data preparation stage, and broadens the application range of the technology in resource limited scenes; a causal reasoning and physical constraint mechanism is introduced, a dynamic relation between targets is modeled through a space-time causal graph, unreasonable tracks are filtered in combination with a physical rule, and complex conditions such as shielding, rapid movement and extreme weather are effectively dealt with; the dynamic feature calibration module corrects feature drift in real time, and ensures stable model performance in long-term tracking.

The invention belongs to the field of explosion load and response prediction, and particularly relates to an impact load and response intelligent prediction system and method based on machine vision. An impact load and response intelligent prediction system based on machine vision comprises an image data preparation module, an image recognition and segmentation module, a numerical calculation module and a load and response prediction module, and the image data preparation module obtains, marks and enhances an image data set required by a target detection and tracking model; the image recognition and segmentation module recognizes and segments splashing fragments and dynamically tracks the splashing fragments; the numerical calculation module generates a sample database required for training the machine learning model; and the load and response prediction module outputs a forward prediction model, a reverse prediction model and a prediction result. According to the prediction system and method, computer vision, a numerical calculation method and machine learning are fused together, and the problems that a traditional method is low in efficiency and cannot conduct reverse prediction according to fragment distribution after impact explosion are solved.

The invention provides a prediction bounding box dynamic correction method based on multi-dimensional kinematics characteristics, and belongs to the technical field of prediction bounding box dynamic correction. The method comprises the following steps: initializing parameters and preparing data; track validity judgment; calculating bounding box offset based on historical displacement offset; calculating track acceleration; detecting a turning state; calculating by combining displacement offset, acceleration and bounding box offset in a turning state; carrying out adaptive threshold calculation based on recent motion trend credibility; adjusting bounding box offset in combination with a self-adaptive threshold value; and adjusting the original prediction bounding box. According to the method, parameters can be dynamically adjusted, extra training data is not needed, the method can serve as a lightweight module to be integrated in an existing tracking or prediction system, bounding box prediction errors can be effectively reduced, the target positioning stability in a complex dynamic scene can be improved, and hardware cost and deployment complexity are saved.

The invention relates to a forensic traumatology deep learning-oriented PMCT image acquisition and multi-modal data processing system and method. The system comprises a trauma model preparation module, a multi-modal data acquisition module, a PMCT image preprocessing module, a multi-scale contrast enhancement image processing module, a multi-modal data fusion module and a deep learning model input data preparation module. According to the method, through standardized trauma model preparation, multi-modal data acquisition, PMCT image preprocessing, multi-scale contrast enhancement processing and multi-modal data fusion, high-quality input data suitable for a deep learning model is constructed, and the problems of detail loss, insufficient information expression and the like when the forensic PMCT image is applied to deep learning are solved. The method has the advantages of standardization, information retention and enhancement, multi-modal fusion, professional knowledge embedding, high adaptability, efficiency improvement, high practical value and the like, and provides more accurate and objective technical support for forensic trauma analysis.

The invention discloses a workpiece surface defect detection method and system, and relates to the field of image processing, and the method comprises the following steps: data preparation: collecting workpiece surface image samples, carrying out sample data processing, and classifying defect images into a micro defect data set and a slender defect data set; model construction: constructing a workpiece surface defect detection network model and a workpiece surface segmentation network model; model training: training a workpiece surface defect detection network model by using the micro defect data set, and training the workpiece surface defect detection network model by using the slender defect data set; and defect detection: respectively inputting a workpiece surface image to be detected into the workpiece surface segmentation network model and the workpiece surface defect detection network model for defect detection. According to the invention, the detection precision is improved, the number of parameters is reduced, and the production efficiency is improved.