556 results about "Diagnostic accuracy" patented technology

The invention provides a self-adaptive bearing fault diagnosis method based on multi-base wavelet fusion. The objective of the invention is to solve the problems of noise reduction, insufficient feature extraction and low diagnosis precision under noise conditions. A Kaisixi University bearing public data set is used as original data, and Gaussian noise with different SNRs is superposed to simulate various noise intensities. And uniformly carrying out length alignment, down-sampling, equal-length segmentation, division and normalization preprocessing. Then, wavelet bases such as sym4, db4, coif5 and the like are adopted for parallel multi-scale decomposition and reconstruction; and adaptively determining the number of decomposition layers and a threshold strategy according to the noise level, and generating a de-noising branch. And performing weighted fusion on the denoising results of the branches, and performing iterative denoising on the residual error. Signals subjected to noise reduction processing are sent to a double-branch convolution-cycle-attention network, a convolution layer extracts features, an LSTM and a self-attention module capture time sequence changes, and accurate recognition of various bearing faults is achieved. The training adopts a segmented attenuation learning rate and an early stop strategy, and the robustness and generalization ability of different SNR working conditions are improved.

The present invention belongs to the technical field of signal processing, and particularly relates to a multi-source data fusion fault diagnosis method and system based on the DS evidence theory. The method of the present invention comprises: first, collecting multi-source information of a mechanical device, wherein the multi-source information includes vibration data, image data and sound data; respectively calculating corresponding feature indicators; then, using a PCA algorithm to perform dimensionality reduction processing on the indicators; using a Bayesian fault diagnosis model to perform fault diagnosis, so as to obtain a diagnosis result from each sensor; and then, using the DS evidence theory to perform fault fusion diagnosis. The present invention can comprehensively consider the characteristics of various types of data, extract more effective fault indicators, achieve more comprehensive fault identification, and fuse diagnosis results from a plurality of sensors, thereby ensuring the accuracy of final diagnosis. The present invention is suitable for fault diagnosis of various complex mechanical systems.

The invention relates to the technical field of transformer fault diagnosis, in particular to a transformer fault diagnosis method and system based on generative acoustic large model enhancement, and the method comprises the steps: constructing a generative acoustic large model, generating specified virtual reference Mel spectrum feature data through an acoustic semantic prompt, and carrying out the recognition of the Mel spectrum feature data; configuring a condition self-encoding model to perform feature reconstruction based on the diagnosis condition label and the current reference Mel spectrum feature, and calculating a reconstruction feature error metric value of the reference Mel spectrum feature and the corresponding reconstruction Mel spectrum feature; comparing the calculated reconstructed feature error metric value with a fault diagnosis error metric threshold value corresponding to the diagnosis condition label; and based on the time sequence characteristics of the reference Mel spectrum features, comparing the fault classification statistical information obtained by statistics with the fault diagnosis statistical judgment conditions to determine the fault diagnosis state of the transformer to be subjected to fault diagnosis. The transformer fault diagnosis accuracy can be improved, and the fault diagnosis cost is greatly reduced.

The invention belongs to the technical field of intelligent operation and maintenance and automatic inspection of power equipment, and particularly discloses a power transformation and distribution station room inspection method and device based on multi-mode visual perception. The method comprises the following steps: controlling the intelligent inspection robot to move and synchronously acquiring visible light, infrared thermal imaging, partial discharge ultrasound and other multi-mode sensing data streams; generating an ultrasonic-guided enhanced infrared image, a power equipment structure map and an ambient light field distribution map through deep fusion in combination with ambient illumination parameters; and on the basis of the deeply fused information, generating a refined health state level of the power equipment, and finally automatically generating an inspection report conforming to the regulation. According to the invention, through deep cooperation and mutual verification of the multi-modal data, the perception robustness and diagnosis accuracy of early faults and abnormal states of equipment in a complex environment are improved, intelligent inspection from later judgment to beforehand prediction is realized, and the power supply safety of a power transformation and distribution station room is guaranteed.

The invention relates to a GIS (Gas Insulated Switchgear) equipment mechanical defect diagnosis method based on a Gramb angle field and a dual-channel PCNN-Attention neural network, and belongs to the technical field of gas insulated switchgear mechanical vibration defect diagnosis. The method solves the problems that traditional diagnosis depends on artificial feature extraction, so that subjectivity is high, information mining is insufficient, and defect severity evaluation is missing. According to the technical scheme, the method comprises the steps that a one-dimensional vibration signal is converted into a GASF two-dimensional image and a GADF two-dimensional image through a GASF field so as to completely reserve time sequence topological features; carrying out data enhancement by adopting an image geometric transformation technology so as to improve the generalization ability of the model; and a dual-channel PCNN-Attention model is constructed, and synchronous intelligent identification of defect types and severity is realized through parallel feature extraction and dynamic weight optimization of an attention mechanism. According to the method, the diagnosis accuracy, reliability and adaptive capacity are improved, and support is provided for equipment state operation and maintenance.

The invention provides an intelligent and rapid pre-examination and diagnosis method and system for tree health, and relates to the technical field of tree management. The method comprises the following steps: constructing a liquid flow database of a standard tree of a target tree species, carrying out model training on an established liquid flow calculation model by utilizing the liquid flow database, then collecting liquid flow data of a to-be-detected target tree in a preset detection time and meteorological and soil information of a growth environment, inputting an optimal liquid flow calculation model, and calculating a theoretical liquid flow value of the target tree, comparing with an actually monitored liquid flow value, calculating a liquid flow deviation degree, grading according to a daily maximum liquid flow deviation degree, and comprehensively evaluating a tree health grade in combination with phenotype information to obtain a diagnosis result; the rapid and intelligent pre-examination diagnosis of the tree health is realized, and the intervention timeliness is improved; and multi-dimensional data fusion is adopted, so that the diagnosis accuracy is improved, and automatic monitoring and real-time early warning are realized.

The invention relates to the technical field of medical training equipment, in particular to a critical care ultrasonic image training system, which comprises a clinical data acquisition module for acquiring critical clinical case data, ultrasonic image materials and diagnosis and treatment process specifications; the training scene construction module is used for constructing a high-fidelity three-dimensional training scene according to severe clinical case data, ultrasonic image materials and diagnosis and treatment process specifications; the virtual ultrasonic practical operation module simulates the operation of an ultrasonic probe based on a high-fidelity three-dimensional training scene, and generates an ultrasonic image corresponding to an anatomical part in real time; the intelligent skill assessment center assesses operation normalization and diagnosis accuracy through practical operation data and real-time generation of an ultrasonic image corresponding to an anatomical site in combination with an improved ResNet-LSTM model, and an assessment result is obtained; and the personalized teaching unit generates a customized training scheme and knowledge strengthening content according to the evaluation result in combination with the learning track of the student. Therefore, the problems of insufficient dynamic scene simulation, poor authenticity and the like in the prior art are solved.

The invention belongs to the technical field of electrical equipment state monitoring, and particularly discloses an electrical equipment intelligent online monitoring system and method based on multi-parameter fusion. Comprising the following steps: synchronously acquiring a partial discharge parameter, a mechanical characteristic parameter and a temperature rise parameter during operation of equipment, performing signal preprocessing on the parameters respectively, and constructing a multi-dimensional feature vector for diagnosis; the multi-dimensional feature vector is combined with a time sequence correlation model, a trend correlation model and an intelligent diagnosis model to output the multi-dimensional feature vector, collaborative decision making is carried out on an analysis result through a D-S evidence theory, and an equipment state evaluation conclusion and early warning information are generated and output in a visual form; according to the method, the monitoring limitation of a single parameter is overcome, early warning can be realized while the diagnosis accuracy is improved, the operation and maintenance efficiency is further improved, and the comprehensive diagnosis after deep analysis is also helpful for operation and maintenance personnel to take corresponding measures in time.

The invention relates to the technical field of medical equipment, in particular to a cardiopulmonary sound and electrocardiogram rapid screening system suitable for emergency treatment scenes. The system comprises a heart and lung sound acquisition unit, an electrocardiogram acquisition unit and an intelligent analysis unit. And the intelligent analysis unit realizes signal synchronization by establishing a cross-modal time sequence alignment model, adaptively matches the physiological state by adopting a dynamic time window adjustment mechanism, and introduces multiple rounds of verification processes to ensure the result reliability. The system can automatically complete synchronous collection, feature extraction and fusion analysis of the heart and lung sound and the electrocardiosignals, the diagnosis accuracy is remarkably improved while the screening speed is guaranteed, and reliable technical support is provided for rapid screening of heart and lung diseases in the emergency department.

The invention provides a skin lesion auxiliary detection method and system based on multi-modal information fusion, and belongs to the technical field of computer vision and application. The method comprises the following steps: firstly, acquiring and constructing bimodal data including a dermatoscope image and a text of a patient, and preprocessing the data; secondly, respectively adopting an improved convolutional neural network and a text coding model, and extracting deep and high-dimensional features from the skin disease image and the self-described text of the patient; afterwards, an improved symmetric gating cross attention fusion module is constructed, fine-grained interaction and alignment between visual and text features are realized through a bidirectional cross attention mechanism, and a gating unit is utilized to adaptively adjust a fusion weight; and finally, inputting the deeply fused multi-modal feature vector into a classifier to obtain a confidence score of each skin lesion type so as to provide reference for dermatologists. According to the method, the diagnosis accuracy, robustness and interpretability are synergistically improved, and the method has great clinical application value.

The invention discloses an intelligent diagnosis method and system based on a large model and knowledge retrieval enhancement, and the method comprises the steps: obtaining multi-source heterogeneous data through a distributed collection network, and building a standardized data flow; extracting multi-dimensional dynamic features to construct a state model, and identifying an abnormal mode; the key innovation lies in that the deep semantic understanding ability of a large language model is combined with a structured knowledge graph, and the exceptions are primarily screened and rechecked through a dual verification mechanism; and finally, closed-loop optimization from diagnosis to treatment is realized. The multi-source data fusion greatly expands the fault perception dimension, the dual verification mechanism of knowledge retrieval and semantic understanding effectively filters false alarms, and the false alarm rate is significantly reduced; a closed-loop feedback mechanism enables the system to have continuous learning ability, and the diagnosis precision is continuously improved along with operation time; the accuracy of fault identification and the timeliness of system response are greatly improved, and technical support is provided for conversion of operation and maintenance of electromechanical equipment from passive response to active intervention.

The invention provides an SPECT pinhole collimator imaging truncation artifact filling method and related equipment, and relates to the technical field of CT imaging. The method comprises the following steps: reconstructing real projection data to obtain an original reconstructed image, and determining a high-confidence region in the original reconstructed image; mapping the real projection data to a projection image of an ideal detector radial distance to obtain mapping projection data; predicting projection data outside an effective area of a real acquisition detector and taking the projection data as simulation projection data; carrying out weighted reconstruction on the simulation projection data to obtain a complemented image containing a truncation region; and according to the high-confidence region, fusing the original reconstructed image and the complemented image to obtain a final image. The method aims at solving the problem of truncation artifacts caused by the fact that the probe cannot completely cover the whole human body projection image, negative effects caused by the truncation artifacts can be made up, and the effect of improving the imaging quality and the diagnosis accuracy is achieved.

The invention discloses a multi-source information fusion fault diagnosis method based on an improved DS evidence theory, and relates to the technical field of industrial equipment state monitoring and intelligent fault diagnosis, and the method comprises the steps: extracting multi-scale vibration energy and impact characteristics, a thermal load change rate and oil physical and chemical characteristics through synchronously collecting vibration, temperature and lubricating oil multi-source operation information; and constructing a unified feature vector and inputting the unified feature vector into the deep belief network to realize initial probability distribution of normal, early warning and fault working conditions. The cross-modal evidence conflict is evaluated and corrected by constructing a feature coupling index, a conflict degree and a confidence entropy index, and the interference of inconsistent evidences on a diagnosis result is inhibited; and further introducing a time sequence evidence library and improving a DS recursive fusion mechanism, depicting time evolution of an equipment operation state, and constructing a health trend index to realize evolutionary fault early warning. And finally, the model is updated in combination with diagnosis result feedback, self-learning optimization and closed-loop fusion of fault diagnosis are realized, and the diagnosis accuracy and stability are improved.

The invention provides a motor intelligent fault diagnosis method and system based on multi-source signal fusion and a self-attention mechanism, and an electronic device. The method comprises the following steps: S1, obtaining a motor multi-source signal; s2, performing multi-scale frequency domain feature extraction and mapping on the motor multi-source signal to obtain a two-dimensional feature matrix; s3, performing spatial dimension feature optimization on the two-dimensional feature matrix by adopting a feature region dynamic enhancement network pre-constructed based on a self-attention mechanism; s4, training a CNN model through the two-dimensional feature matrix after spatial dimension feature optimization; s5, inputting the data of the motor to be tested into the trained CNN model, and diagnosing the fault of the motor to be tested; the method achieves the recognition and enhancement of a fault sensitive feature region in a single time frame, inhibits the interference influence of irrelevant features, and enables the diagnosis precision of a CNN model to be higher.

The present invention relates to an AI-based system and method for generating enhanced radiology reports. The system comprises a database for storing multimodal patient data, a natural language processing (NLP) module for extracting clinical information, and a machine learning module for correlating the clinical information with radiology images to identify diagnostic insights. An AI-based report generation module analyzes the images and clinical information to generate a preliminary report, which is refined based on radiologist input. The generated report is then integrated into the patient's electronic health record. The system employs techniques such as multimodal deep learning, active learning, explainable AI, and federated learning to enhance diagnostic accuracy, capture expert feedback, provide transparency, and enable multi-institutional collaboration. The invention aims to improve the accuracy, efficiency, and value of radiology reporting in patient care.

The invention provides a double-current comparison and DHI combined equipment fault detection method and system, and belongs to the technical field of power equipment state monitoring and fault diagnosis. The method comprises a time sequence perception adversarial enhancement module, a generative adversarial network (GAN) data expansion module, a double-flow contrast attention network (D-CAN) feature extraction module and a dynamic health index (DHI) calculation module. The core lies in that a fault sample with time sequence correlation is supplemented through a physically constrained GAN, robust features are extracted by using a ResNet1D and Transform fused double-flow network, and a health state is quantified in combination with unsupervised clustering and mahalanobis distance. Through simulation and experimental verification, zero-delay detection of the early fault of the transformer can be realized, the output health index and the 3D visualization result can provide an accurate basis for operation and maintenance of the transformer, and the diagnosis accuracy, the data utilization rate and the dynamic adaptability of state evaluation are remarkably improved.

The invention discloses a motor bearing fault diagnosis method and system based on multi-modal fusion and few-sample learning, and relates to the field of motor bearing fault diagnosis, and the method comprises the steps: constructing a multi-modal time sequence signal sample set and a motor bearing fault diagnosis model based on multi-modal fusion and few-sample learning; performing Riemannian metric-based quality weighted element training on the motor bearing fault diagnosis model by using the multi-modal time sequence signal sample set; performing rapid fine adjustment on the trained motor bearing fault diagnosis model by using a small number of marked multi-mode time sequence signals of the target motor bearing; and performing fault diagnosis on the to-be-detected multi-mode time sequence signal of the target motor bearing by using the finely-adjusted motor bearing fault diagnosis model. The method can effectively improve the diagnosis accuracy and robustness.

Disclosed in the present invention are a fault diagnosis method and system for blade icing and blade mass imbalance of wind turbines. The method comprises the steps of: acquiring operation data and blade icing information of a plurality of wind turbines; labeling the operation data with an icing state label on the basis of the blade icing information, so as to obtain fault data; extracting fault features in the fault data, and ranking the fault features according to the degree of importance, so as to generate an optimal feature set; on the basis of a criterion of minimizing a squared error, selecting optimal features in the optimal feature set to generate an optimal decision tree; and performing classification on the basis of the optimal decision tree, so as to obtain a diagnosis test result including fault information. The present invention has the advantages of a high level of diagnostic accuracy, etc.

The invention relates to an artificial intelligence diagnosis auxiliary method and device based on a medical image, equipment and a medium. According to the method, standardized images and interested area masks are extracted from medical images to serve as basic data, in combination with a pathological causal atlas matrix constructed by medical domain knowledge, the atlas matrix is utilized to guide an attention mechanism in a deep neural network to generate a causal-associated weighted feature map and attention distribution; further eliminating the influence of confusion variables through adversarial training and causal intervention loss processing so as to obtain a robust diagnosis model, and finally performing path search and confidence calculation on attention distribution and a pathological causal map based on an analysis result of the model on a target image. And a diagnosis decision path and a visual interpretation report conforming to clinical causal logic are generated, so that the false correlation feature interference is effectively inhibited while the diagnosis accuracy is ensured, and the transparency and clinical credibility of a model decision process are remarkably improved.

The invention relates to the technical field of wastewater pipelines, in particular to a health state diagnosis method for an electrochemical energy storage prefabricated cabin wastewater collection pipeline, which comprises the following steps: collecting an operation signal of the pipeline, simulating the operation state of the pipeline based on a preset simulation model, obtaining an operation virtual signal of the pipeline, comparing the operation signal with the operation virtual signal, and determining the health state of the pipeline. Operation deviation is obtained; extracting the characteristics of the operation deviation to obtain a deviation characteristic vector; and inputting the deviation feature vector into a fault classification model obtained by training, identifying a potential abnormal type of the pipeline, and outputting a health state index. By collecting flow, sound and vibration signals, generating operation virtual signals in combination with a simulation model, comparing and analyzing deviation characteristics, inputting the deviation characteristics into a fault classification model, identifying potential anomalies and outputting health state indexes, quantitative evaluation and early warning of pipeline health are achieved, the pipeline operation state can be reflected in multiple dimensions, and the reliability of the system is improved. Diagnosis accuracy and system reliability are improved, and health grade division and operation and maintenance decision making are supported.

The invention provides a dynamic knowledge graph construction and diagnosis reasoning method oriented to intelligent inquiry, and relates to the technical field of knowledge graphs, comprising the following steps: acquiring multi-source medical data, performing entity recognition and semantic annotation, establishing a causal probability graph based on a structured entity set, and establishing a dynamic knowledge graph; and selecting an optimal questioning problem according to the information gain in the inquiry process, dynamically updating the causal probability by using the Bayesian rule, and finally propagating the conditional probability along the causal path to generate a diagnosis conclusion. According to the invention, personalized inquiry decision and accurate diagnosis reasoning are realized, and the diagnosis accuracy and efficiency of the intelligent inquiry system are improved.

The invention discloses an allergen specific IgE and IgG antibody composite quality control product or calibration product as well as a preparation method and application thereof, and belongs to the technical field of biological detection. The preparation method comprises the following steps: obtaining a whole-genome sequence comprising a human IgE constant region gene sequence, a human IgG4 constant region gene sequence and a human IgG1 constant region gene sequence, and preparing a fusion protein simultaneously containing a human IgE Fc segment and a human IgG Fc segment; immunizing a healthy animal by using the allergen to obtain a purified antibody; and carrying out chemical coupling on the purified antibody and the fusion protein to obtain allergen specific IgE and IgG antibodies. By constructing allergen specific IgE and IgG antibodies, the kit can be simultaneously applied to detection of allergens IgE and IgG, has the advantages of high titer, high purity, good dilution linearity and the like, and is beneficial to improvement of diagnosis accuracy and treatment effectiveness.

The invention relates to the technical field of visual function diagnosis, and provides a visual function examination device based on eye movement tracking, a medium and computer equipment, the device comprises a data acquisition unit, an eye movement evaluation unit, a visual field evaluation unit and a comprehensive analysis unit, the data acquisition unit is used for acquiring visual function data of an examinee, and the comprehensive analysis unit is used for analyzing the visual function data of the examinee; the eye movement evaluation unit evaluates an eye movement function based on eye movement parameters in the visual function data; the visual field evaluation unit evaluates a visual field function based on visual field parameters in the visual function data; according to the application, objective sight drop points are used for replacing subjective keys, false negative and false positive are reduced, before typical defects do not appear in the visual field, early visual function changes can be prompted through eye movement parameter abnormity, auxiliary screening of early glaucoma and other diseases is achieved, and the visual function detection accuracy is improved. Meanwhile, the diagnosis accuracy is improved through double evidences of eye movement and visual field in the middle and late stages, and the sensitivity and specificity of visual function examination are remarkably improved on the whole.

The invention discloses a well mining unmanned vehicle fault diagnosis method, and relates to the field of unmanned driving, and the method comprises the steps: collecting the multi-sensor data of a well mining unmanned vehicle, and the multi-sensor data comprise vehicle hardware state data, system operation state data and environment perception data; preprocessing the multi-sensor data to obtain a multi-source data set; according to the multi-source data set, the vehicle state is analyzed through a fault diagnosis algorithm, and potential faults are obtained; grading the potential faults to obtain fault grades; and executing a corresponding fault processing strategy according to the fault level. In order to solve the problem of low vehicle fault diagnosis precision caused by the fact that sensor data is easily interfered by environmental factors in a well mining environment in the prior art, the fault source accurate positioning capability and diagnosis accuracy are improved by constructing a multi-level rule base and a fault causal relationship network.

The invention discloses a state diagnosis and early warning system and method for power switch equipment, and belongs to the technical field of power equipment monitoring. According to the system, multi-dimensional sensing data of temperature, vibration, partial discharge and the like of the switch cabinet are uniformly collected and locally processed through an edge calculation module of a built-in multi-protocol Internet of Things gateway. On one hand, the edge calculation module carries out real-time state judgment based on a multi-dimensional threshold value; and on the other hand, the trend diagnosis module analyzes historical data by using an XGBoost algorithm to realize trend prediction and early anomaly recognition. And the comprehensive diagnosis module fuses the two types of results and dynamically adjusts the state level so as to realize accurate early warning. According to the method, comprehensive perception and on-site intelligent analysis are realized, the problems of one-sided perception, response delay, insufficient intelligence and data island existing in a traditional scheme are effectively solved, and the real-time performance, diagnosis accuracy and operation and maintenance efficiency of equipment state monitoring are remarkably improved.

The invention relates to the technical field of multi-source heterogeneous data analysis, in particular to a multi-source heterogeneous data-oriented adaptive analysis model construction method and system, and the method comprises the steps: collecting a multi-source heterogeneous data family of a plurality of bolts, and screening alternative diagnosis data to generate a multi-modal key feature vector; judging whether the multi-modal key feature vector is abnormal or not based on the multi-modal key feature vector and the clustering feature vector; obtaining an abnormal multi-modal key feature vector, and generating a fusion feature vector based on a deep learning model; generating an initial diagnosis result based on the fused feature vector; comparing the diagnosis accuracy with a preset threshold value, and judging whether to trigger deep learning model retraining or not; and correcting the attention center of gravity based on the error feature vector and the clustering feature vector, and generating a confidence diagnosis result to position abnormal dimensions and abnormal data. According to the method, the multi-source heterogeneous data analysis efficiency and the anomaly detection accuracy in the bolt tightening process of the battery pack production line are improved.

The invention relates to the technical field of traditional Chinese medicine intelligent diagnosis, in particular to a pulse condition feature-based traditional Chinese medicine intelligent diagnosis system, which comprises a pulse signal acquisition module, a depth feature extraction and characterization module, a pulse condition semantic mapping and identification module, a personalized adaptation and robustness enhancement module and a diagnosis report module. According to the invention, comprehensive, accurate and robust analysis of pulse condition information is realized, and the accuracy of traditional Chinese medicine intelligent diagnosis is improved.

The invention discloses a power grid monitoring alarm event handling method and system based on a large model, and relates to the technical field of power systems, and the method comprises the steps: obtaining a historical sample of a power grid monitoring alarm event, and carrying out the preprocessing; constructing a balance training sample set based on a first large language model and resampling combined sample enhancement method; based on the pre-training language model, utilizing the balance training sample set to carry out fine tuning training, and constructing a power grid monitoring alarm event diagnosis model; on the basis of a second large language model, a power grid monitoring alarm event disposal model is constructed, and the disposal model is used for retrieving reference information from a historical disposal knowledge base in combination with an event type diagnosis result output by the diagnosis model and generating an auxiliary disposal suggestion for an input alarm event text; and integrating the diagnosis model and the disposal model to form a power grid monitoring alarm event disposal framework. According to the invention, intelligent diagnosis and auxiliary disposal of the power grid monitoring alarm are realized, and the diagnosis accuracy and the disposal efficiency are improved.

The invention discloses a zero sample industrial process fault diagnosis method based on hybrid attribute removing topology and multi-label graph convolution. Firstly, a class-attribute matrix is constructed according to engineering semantics, and a sample is standardized; secondly, training an independent attribute predictor and measuring attribute dependence by using a residual error to obtain an attribute coupling adjacent matrix; using the matrix to drive the GCN to obtain attribute structure embedding, performing multi-label discrimination with sample feature fusion, and adopting loss of asymmetric focus weighting, negative probability cutting and dynamic weight adjustment; and finally, inference is completed by using the predicted attribute vector and the nearest neighbor of the unseen class prototype. The method inhibits spurious correlation caused by shared input, and improves the accuracy and stability of zero sample diagnosis.

The invention provides a power transmission cable insulation state evaluation method based on multi-parameter fusion, which relates to the technical field of cable insulation evaluation, and comprises the following steps: constructing a cable topological graph, dividing the whole cable into a plurality of partitioned cables, and marking environmental data for the partitioned cables; testing voltage is applied to each partitioned cable, multi-parameter data are collected, the multi-parameter data are preprocessed, feature data are extracted, and reference weight coefficients of the feature data of each partitioned cable are obtained; acquiring an environment factor according to the environment data of each partitioned cable; determining a target weight coefficient set in different extreme environments, and obtaining a dynamic weight coefficient of each piece of feature data according to the reference weight coefficient, the target weight coefficient and the environment factor; and performing weighted synthesis on each feature data of each partitioned cable and the corresponding dynamic weight coefficient to obtain an insulation health index of each partitioned cable. A multi-parameter information fusion technology is adopted, the one-sidedness of single-parameter evaluation is overcome, and the diagnosis accuracy is improved.