18653 results about "Training set" patented technology

The present invention discloses an aero engine bearing fault diagnosis method based on STFT-IncepNext, which includes the following steps: step 1: collect acceleration signals under various working conditions and fault conditions of the bearing as sample data; step 2: determine the number of sampling points of each sensor first, and perform data merging by permuting and combining data collected by sensors at different positions in the same time series, and then use short-time Fourier transform to convert the fused vibration signal into a time-frequency diagram; step 3: data preprocessing: divide into training set, test set and validation set according to fault type, and normalize the converted time-frequency diagram using average value and standard difference; step 4: build an IncepNext model, train and validate it; and step 5: use the trained IncepNext model to diagnose aero engine bearing faults and obtain a diagnosis result. The present invention can effectively identify aero engine bearing faults, can guarantee the fault identification rate, and has good accuracy.

A system and method for machine learning-based data field validation is proposed that utilizes a specific trained machine learning model data architecture that is adapted to be more resilient against training set class imbalance, using a Siamese triplet LSTM network architecture that uses three LSTMs that are trained together and operate in concert. An example non-limiting practical use includes using the Siamese triplet LSTM network architecture to validate whether free-text data fields include a single jurisdiction in an address or multiple jurisdictions in the address.

The invention relates to an intelligent question-answering system optimization method and device based on a knowledge graph, and the method comprises the steps: obtaining original knowledge data of a target knowledge domain, and constructing a knowledge graph structure model; extracting term information of entity nodes in the knowledge graph structure model, and constructing an entity term set; receiving a natural language question input by a user, executing a semantic understanding operation based on the standardized expression set to obtain a structured question semantic representation, and matching the question semantic representation with the case training set to obtain context semantic features; constructing a cue word template, and executing a query instruction generation operation to obtain a target query statement of the graph database; submitting the target query statement to a graph database to execute data retrieval operation, and obtaining query result data corresponding to the question semantic representation; and performing personalized rendering processing on the query result data based on the user portrait information to generate final question and answer return content. The method has the effect of improving the query accuracy.

The present disclosure relates to the technical field of target detection, and particularly relates to a model deployment method, an end-side device and a storage medium, which are used for solving the problem in the related art of the accuracy of a deployed model being low. The method comprises: performing target detection on a video frame image input into a first model, and acquiring a first target detection result and a first confidence; if the first confidence is greater than or equal to a first confidence threshold value, recording the video frame image and the first target detection result as samples in a training set; if the first confidence is less than the first confidence threshold value, performing target detection on the video frame image on the basis of a second model, and recording the video frame image and an acquired second target detection result as samples in the training set; and training the first model on the basis of the training set, and replacing the current first model with a trained first model for subsequent target detection. In this way, the accuracy and model generalization capability of a first model are improved.

A target detection method, system and device based on cross-modal fusion and an attention guiding mechanism and a medium, the target detection method comprises the following steps: obtaining a visible light-infrared image pairing data set, and carrying out data processing and division on the visible light-infrared image pairing data set to obtain a training set, a verification set and a test set; constructing a multi-modal target detection network; setting network training parameters; training and optimizing the multi-modal target detection network by using the training set, outputting a training weight file after the training is finished, verifying the trained weight file by using the verification set, and selecting the weight file with the highest precision as an optimal weight file; loading the test set and the optimal weight file into a multi-modal target detection network, and performing target detection on the test set to obtain a target detection result; the system, the device and the medium are used for bearing and implementing the method. The method has lower false detection and false detection conditions, and improves the robustness and accuracy of target detection.

The invention discloses a power load prediction method based on multi-source heterogeneous data feature fusion, system equipment and a medium, and relates to the technical field of power system load prediction, and the method comprises the steps: obtaining a power load prediction data set, and carrying out the preprocessing; performing input feature processing on the static data, the graph data and the time series data based on a multi-layer perceptron, a graph convolutional network and a gating loop unit network, and fusing the processed features based on a feature splicing method to obtain comprehensive feature information; based on full-connection neural network construction, taking the comprehensive features as input, inputting training set data into a model, and performing iterative optimization according to a model effect to adjust model parameters and hyper-parameters; and based on the parameter optimization adjustment model, using the evaluated optimal model to predict the power load in the future time period. The method is superior to a traditional method in precision and stability, can better adapt to complex and changeable load prediction scenes, and has wide application prospects and remarkable actual value.

The invention discloses a wind turbine generator gearbox fault diagnosis method and system based on time-frequency feature cross complementation and multi-domain feature fusion. Collected original vibration signals are divided into a training set, a verification set and a test set; respectively utilizing FFT and CWT to obtain a one-dimensional frequency domain signal and a two-dimensional time-frequency graph of the data set; inputting the training set of the time domain signal and the frequency domain signal into a time-frequency cross attention mechanism network to obtain cross complementation characteristics of the time domain and the frequency domain; inputting the training set of the two-dimensional time-frequency graph into a two-dimensional residual CNN into which a complete attention mechanism is introduced to extract time-frequency domain features; fusing the cross complementation features of the time domain and the frequency domain and the time-frequency domain features to obtain multi-domain features, and performing fault classification by using a softmax function; and inputting test set data into the trained multi-domain feature fusion network model for testing to obtain a final fault diagnosis result. According to the method, the problem of incomplete feature extraction of deep learning fault diagnosis based on a single time domain signal and a frequency domain signal is solved, and the fault diagnosis precision is improved.

The invention discloses a remote sensing image building extraction method fusing double-space attention features, and belongs to the technical field of image processing. The method comprises the following steps: firstly, inputting a training set of an Inria aviation data set into a ConvNeXt backbone network to extract shallow feature information; then, carrying out feature information enhancement processing on the feature information through a double-space attention feature fusion module (DAFF); and finally, a multi-modal dynamic enhancement module (FAM) is used in the last time of feature information up-sampling operation to carry out channel alignment operation of different levels of features, and feature information fusion errors are reduced. According to the method, the accuracy of building extraction in the remote sensing image is remarkably improved.

The invention provides an ultrahigh-resolution remote sensing image segmentation method based on frequency domain information fusion. The ultrahigh-resolution remote sensing image segmentation method based on frequency domain information fusion comprises the following steps: step 1, acquiring a Vaihingen and Potsdam remote sensing data set; step 2, building an ultrahigh resolution remote sensing image segmentation network fused with multivariate frequency domain information; step 3, inputting the Vaihinge training set into the ultra-high resolution remote sensing image segmentation network for training to obtain an optimal parameter model; 4, inputting the Vaihinge test set into the optimal parameter model, and outputting an accurate segmentation image of the remote sensing image; and 5, replacing the data set with a Potsdam training set, a Potsdam verification set and a Potsdam test set, and repeating the operations from the step 3 to the step 4. According to the model, the LKResNet-18 network is introduced, so that the receptive field and the feature representation capability of local branches are improved.

The invention discloses a point cloud semantic segmentation method and system based on local neighborhood attention, and belongs to the technical field of three-dimensional point cloud processing and deep learning, and the method comprises the following steps: S101, dividing three-dimensional point cloud data into training, verification and test sets, and carrying out the preprocessing; s102, constructing a deep learning network based on local neighborhood attention, training by using the training set, and obtaining an optimal model by using the verification set; and S103, inputting the test set into the optimal model to obtain a semantic segmentation result. According to the method, the semantic segmentation model is constructed by fusing local neighborhood attention of global position coding and local position coding, so that the problem of poor space geometric structure information extraction due to the fact that a conventional model only pays attention to the relative position of points in a neighborhood is effectively solved, and the feature extraction capability of an attention mechanism is enhanced; and the robustness, generalization and segmentation accuracy of the three-dimensional point cloud semantic segmentation model are improved.

The invention discloses a traffic flow prediction method based on graph diffusion and dynamic graph fusion. The method comprises the following steps: S1, acquiring historical traffic flow time sequence data of each traffic node in a target road network; s2, preprocessing historical traffic flow time series data to obtain a road network node adjacency matrix; taking the historical traffic flow time sequence data and the road network node adjacency matrix as sample data, and dividing a training set, a verification set and a test set according to a preset proportion; s3, constructing a traffic flow prediction model based on graph diffusion and dynamic graph fusion; and S4, performing model training and verification on the traffic flow prediction model through the training set and the verification set to obtain an optimal traffic flow prediction model, and realizing traffic flow prediction of the test set through the optimal traffic flow prediction model. The problems that an existing method does not have the dynamic topology modeling capacity, the high heterogeneous feature fusion capacity and the self-adaptive space-time modeling capacity, and consequently the bottleneck problem of a current model in the aspects of prediction precision, stability and practicability cannot be effectively solved.

The invention relates to an automatic driving large model training optimization method based on multi-scene data balance. Comprising the following steps: (1) constructing a real vehicle high-speed driving scene library; (2) constructing a visual language automatic driving large model, training by adopting an iterative training framework based on a real vehicle high-speed driving scene library, designing a multi-task joint loss function and a weight adaptive adjustment strategy, realizing multi-task target balance, and obtaining a trained visual language automatic driving large model; (3) dynamic simulation is carried out for the automatic driving working condition, and a high-fidelity simulation data test set is constructed based on simulation data; and (4) performing hyper-parameter optimization and lightweight processing on the trained visual language automatic driving large model according to the high-fidelity simulation data test set to obtain a scene data balanced automatic driving large model. According to the method, the large model reasoning speed is increased, and resource occupation is reduced, so that the judgment capability of the large model on dynamic working conditions and high-risk scenes is remarkably improved.

The invention provides a multi-modal large model construction and operation method and system for reservoir dam safety. The method comprises the following steps: constructing a training set, wherein the training set comprises a single sample; multi-modal data as model input: sensor data arranged on the dam, historical hidden danger records, engineering technology documents formed in various stages of planning, design, construction and operation of the dam, information resources related to dam safety, geometric and material information of the BIM model, timestamps and spatial position information; the probability value of occurrence of the hidden danger event serving as the training label is obtained; the training set is adopted to train a multi-mode pre-training model based on a Transform framework; performing real-time prediction and early warning on potential safety hazards of the dam; and calculating the actual influence degree, the comprehensive risk score, the dynamically set early warning threshold and the emergency measure suggestion of the hidden danger event according to the probability value of whether the hidden danger event occurs or not output by the model. The scientificity and effectiveness of dam safety monitoring are comprehensively improved.

The invention provides a deep learning-based avalanche early warning model construction method and system, and the method comprises the steps: firstly obtaining multi-source environment monitoring data, including meteorological time sequence, topographic space and accumulated snow layer physical data, of a target region, carrying out the time dimension alignment of the meteorological time sequence data to generate a feature sequence, carrying out the meshing of the topographic space data to generate a feature set, and carrying out the construction of an avalanche early warning model; the method comprises the following steps: extracting parameters from accumulated snow layer physical data to generate a state vector, inputting a deep learning network model containing time sequence attention, spatial convolution and cross-modal interaction units, generating a fusion feature vector, constructing a training set based on historical avalanche event annotation data, performing dynamic weight optimization on the fusion feature vector, and generating an avalanche risk prediction model. And finally, receiving current monitoring data in real time, outputting a risk level and an early warning trigger threshold value by the avalanche risk prediction model, and generating a multi-level early warning signal when a real-time risk value exceeds the threshold value, thereby realizing accurate avalanche early warning.

The invention discloses an electric power system abnormal remote signaling detection method based on a graph auto-encoder model, and the method comprises the steps: collecting measurement data of an electric power system, carrying out the preprocessing, obtaining a graph data set with an abnormal label, and dividing the graph data set into a training set and a test set; inputting the graph data in the training set into the graph auto-encoder model for training; after training is completed, abnormal score distribution is counted based on normal edge samples in a training set, a threshold value is set to serve as a follow-up judgment basis, and threshold value selection takes the accuracy rate and the recall rate on a test set as an adjustment and optimization target; in a test stage, image data in a test set are input to carry out edge feature reconstruction and anomaly scoring, anomaly judgment is carried out on edges in combination with a set threshold value, and a preliminary abnormal edge detection result is output; and the output abnormal edge detection result is input into the graph restoration module, the restored edge structure and edge features are output, the damaged remote signaling state in the power grid is restored, and the integrity of the graph structure and the operation credibility of the power system are improved.

The invention discloses a deep foundation pit stability deformation control method and device based on digital twinning. The method comprises the steps that monitoring time sequence data of a deep foundation pit are collected and obtained; constructing a deep foundation pit deformation control digital twin platform; constructing a CNN-LSTM composite neural network based on an Adam optimization algorithm through a foundation pit deformation prediction module, and performing model training / verification on the CNN-LSTM composite neural network through the data training set and the data verification set to obtain an optimized CNN-LSTM composite neural network; based on the optimized CNN-LSTM composite neural network, performing time sequence change prediction on the axial force of the steel support according to the data test set to obtain a prediction result; the steel support servo module realizes servo control on the axial force of the steel support according to a prediction result on the basis of a constructed servo control strategy; and the foundation pit BIM module is used for constructing a deep foundation pit BIM model based on a Revit modeling technology and interacting with the deep foundation pit BIM model in real time through a RevitAPI data interface according to the output of the steel support servo module so as to realize visualization of deformation control of the deep foundation pit. The problems that in the deep foundation pit construction process, due to the fact that complex geological conditions are involved, the surrounding environment and an existing building are affected to a certain degree, and deformation and stability of the deep foundation pit cannot be accurately controlled are solved.

The invention discloses a deep reinforcement learning-based network traffic optimization scheduling method suitable for periodic traffic characteristics. The method comprises the following stages: collecting historical periodic traffic data; a network state is monitored in real time through an SDN controller, and historical periodic flow data including time delay, bandwidth utilization rate and load information are acquired; data preprocessing: performing cleaning, normalization and time sequence segmentation on the collected historical traffic data; constructing an LSTM model, constructing the LSTM model by using the preprocessed historical data, and defining an input feature and an output tag; training an LSTM model, training the LSTM model by using the training set data, and generating traffic prediction data of a next period; constructing a state space, fusing the predicted traffic data with the real-time traffic data, and constructing the state space; training a DQN model, and training a deep Q network (DQN) model by using the constructed state space; network topology and link state information are obtained, the current network topology is obtained through a topology discovery module of the SDN controller, and the bandwidth utilization rate, time delay and packet loss probability information of a link are collected in real time through a flow monitoring module; large flows and small flows are distinguished; scheduling the small flow by using the ECMP; and scheduling the large flow by using the DQN model.

A knowledge distillation method and system based on embedded feature regularization. The method comprises: collecting labeled image data related to a recognition task; calculating a feature center of a teacher model in each category on a whole training set; projecting in the direction of a category center an embedded feature of a student model, rotating to the direction of the category center an embedded feature of the teacher model, and constructing a feature regularization loss by means of regularizing a projection feature of the student model and a rotation feature of the teacher model; increasing the norm of the feature of the student model, and constraining the direction of the feature of the student model to be consistent with the direction of the category center; inserting the feature regularization loss into an existing knowledge distillation framework, and training the student model; and deploying the trained student model to a terminal device, predicting a logits vector on the basis of new data received by the terminal device, and thus completing the related recognition task. In the present invention, by means of increasing the norm of a feature of a student and constraining the direction of the feature to be consistent with the direction of a category center, the knowledge distillation performance is better.

The invention relates to the technical field of computer vision detection, in particular to an unmanned aerial vehicle target detection method based on frequency-space joint attention and dynamic fusion, and the method comprises the steps: obtaining an unmanned aerial vehicle image data set, carrying out the preprocessing, and dividing a training set and a test set; constructing a target detection model, inputting the training set into the target detection model to extract image features, sequentially performing frequency domain detail enhancement, spatial domain salient region extraction and multi-scale feature adaptive fusion based on the image features, and establishing a feature sequence; screening the feature sequence to obtain an initial target query, and finishing target classification and positioning on the initial target query through a decoder; training a target detection model by using the training set, and inputting the test set into the trained target detection model to generate a detection result; on the premise that the real-time reasoning advantage of RT-DETR is kept as much as possible, the problems that in an unmanned aerial vehicle scene, a target is prone to missing detection, the scale change is large, the background is complex, and the target is fuzzy are effectively solved, and the detection precision is improved.

The invention provides a clamping force correction method and system based on an oil-tea tree branch clamping manipulator and a storage medium, and relates to the technical field of mechanization and automation, and the method comprises the steps of monitoring and analyzing oil-tea tree branches, an operation environment and a manipulator motion state, collecting key parameters, generating a clamping adaptability index and a clamping stability index, and correcting the clamping force of the oil-tea tree branches; the clamping adaptability index and the clamping stability index are combined, and a clamping control index is calculated; constructing a deep learning model, taking the historical key parameters as a training set, taking the corresponding historical manipulator clamping force as a label, training the model to obtain a clamping force prediction model, inputting the key parameters collected in real time into the model, outputting the predicted clamping force, combining the predicted clamping force with a clamping control index, and calculating the final clamping force. On the basis of obtaining key parameters in real time, the clamping adaptability index and the clamping stability index are generated, then the clamping control index is calculated, and the purpose of dynamically adjusting the clamping force is achieved.

The invention discloses a road defect detection method based on heavy parameter multi-scale fusion, and belongs to the technical field of target detection in a deep learning neural network, and the method comprises the steps: selecting a data set of road defects from a public data source, and dividing the data set into a training set, a verification set and a test set; a lightweight road defect detection neural network model based on heavy parameter multi-scale fusion is constructed, the model adopts a heavy parameterization technology and a multi-scale feature fusion strategy, and various road defects in the image can be identified and positioned; the constructed neural network model is trained by using the divided training set and verification set, an optimal lightweight road defect detection model is obtained through index evaluation, then reasoning acceleration of the lightweight detection model is carried out, and the lightweight detection model is embedded into an edge computing terminal for deployment; a vehicle-mounted camera is used for collecting image data, collected road images are input into an edge computing terminal to be processed, and types and positions of road defects existing in the vehicle-mounted images are output.

The invention provides an intelligent data quality monitoring method and system. The method comprises the following steps: performing data processing on multi-source heterogeneous data by utilizing a unified data model; inputting historical data as a training set into the pre-training model for training optimization to obtain an optimized pre-training model, and performing threshold determination on the standardized distribution data stream by using the optimized pre-training model in combination with an adaptive threshold module; inputting the processed abnormal event list into a context awareness and dynamic priority management module to carry out de-duplication optimization transmission; carrying out root analysis on the abnormal event list after the de-weighting optimization by utilizing a root analysis tool; generating a repair strategy for the root analysis result by using the assistance of a repair suggestion engine; and inputting the repair strategy and the visual report into a cross-platform monitoring interface, and generating a repair suggestion and an early warning notification. According to the method, the data consanguinity map is constructed based on the graph database, minute-level traceability of abnormal events is achieved, and the troubleshooting time is shortened by 50%.

The invention discloses a double-flow remote sensing image change detection method fused with Mama enhancement, which belongs to the field of image processing and comprises the following steps: acquiring images to obtain a remote sensing image data set; preprocessing the obtained dual-time-phase remote sensing image, and dividing the image into a training set and a test set; designing a dual-flow change detection network model integrated with Mama enhancement; training the constructed double-flow change detection network model by adopting training set data until the whole model is converged, and storing an optimal model; and inputting test set data into the trained optimal model, and predicting a change area in the test set. According to the method, MambaBlock and a semantic segmentation aggregation module are introduced, the global feature modeling capability is enhanced under linear complexity, and the detection precision and robustness of a change region are improved by fusing feature information of different levels and multi-scale feature learning.

The invention relates to the field of remote sensing image processing, in particular to a high-resolution remote sensing image semantic segmentation method based on multi-scale feature fusion, which comprises the following steps: acquiring a public remote sensing image data set, preprocessing the image, and constructing a training and testing set of semantic segmentation; a CTMFNet is designed, an encoder is composed of a lightweight residual module and an MS-Transform, and local space details and global context information are extracted; rID is adopted to reduce spatial information loss, LSFE is introduced to improve spatial positioning capability, and feature calibration is carried out in space and channel dimensions through DecoderAttn to realize boundary fine segmentation; inputting the training sample into the network for training to obtain a converged optimal semantic segmentation model; and inputting the test set into the model to obtain a semantic prediction map, and outputting a fine segmentation result of the remote sensing image through multi-scale fusion and boundary restoration. According to the method, the precision and robustness of ground feature extraction are effectively improved, the calculation cost is remarkably reduced while high segmentation precision is kept, and the method has good practical value and popularization prospects.

The invention belongs to the technical field of intelligent forecasting of ship performance, and discloses a ship resistance forecasting method and system based on a physical information neural network. The method comprises the following steps: carrying out numerical simulation on the direct navigation motion of a ship in still water to obtain a pressure field and a speed field of a watershed and ship resistance information during the direct navigation motion; constructing and forming a total data set, and establishing a training set, a test set and a verification set; establishing a PINN neural network model fused with physical knowledge; obtaining a neural network model based on physical information; generating a random forest model and training; and performing bidirectional verification on the neural network model based on the physical information by using a random forest model to form a network model for ship resistance forecasting. According to the method, the reliability of the forecasting result in the physical level is effectively improved, and the training efficiency of the neural network model is improved, so that the forecasting efficiency is improved.

The invention provides a vehicle driving track prediction method, system and device and a medium, and belongs to the technical field of vehicle track prediction.The method comprises the steps that vehicle state information and environment information of each track point of a vehicle are collected to serve as track data, and a deep learning denoising algorithm is used for preprocessing and feature sequence extraction; constructing a trajectory prediction model by using a Transform in combination with a graph neural network, constructing a training set by using the extracted feature sequence to train the trajectory prediction model, and adjusting model parameters by using a multi-task loss function constructed by trajectory prediction precision and uncertainty estimation in the training process; and using the trajectory prediction model to predict the vehicle trajectory, using the Bayesian optimization method to optimize the trajectory parameters with the goal of minimizing the prediction error, and adjusting the optimization parameters with the vehicle dynamics constraint and the road constraint as constraint conditions in the optimization process. According to the invention, through multi-fusion and multi-stage optimization, high-precision prediction of the vehicle track in a complex traffic environment is realized.

The invention provides an unsupervised industrial anomaly detection method and system based on knowledge distillation, and the modeling method comprises the steps: obtaining an industrial product surface image, and carrying out the preprocessing of the industrial product surface image, so as to obtain a training set and a test set; constructing an abnormal synthesis network, a teacher feature extraction network, a variable de-noising student network and a pixel-level abnormal segmentation network, and integrating the abnormal synthesis network, the teacher feature extraction network, the variable de-noising student network and the pixel-level abnormal segmentation network in sequence to form an abnormal detection model; training the anomaly detection model through the training set; verifying the trained anomaly detection model by adopting the test set, and outputting an anomaly detection target model after verification meets the requirement; according to the invention, the sensitivity and accuracy of industrial image anomaly detection are improved, and more reliable detection, identification and positioning of product surface defects are realized.

The invention relates to a multi-source remote sensing image classification method based on a key band retrieval attention mechanism, and belongs to the technical field of remote sensing image processing. The method comprises the steps of firstly performing preprocessing and data set division on multi-source remote sensing data, then constructing a key band retrieval attention network, performing training and evaluation on a model by utilizing a training set and a verification set after division, and finally performing visual analysis on a model result by utilizing a test set. According to the method, the features of the hyperspectral image and the laser radar / synthetic aperture radar image are effectively extracted and fused, redundant information interference is effectively reduced, the retention rate of hyperspectral key information is improved, the complementary expression ability among multi-source data is enhanced, and the remote sensing image classification precision and calculation efficiency are remarkably improved. The method is suitable for application scenes of multi-source remote sensing data fusion and classification, can meet efficient intelligent processing requirements of complex earth surface information, and provides an accurate and efficient remote sensing image classification solution.

The invention discloses a hyperspectral image classification method based on proxy attention and a VIT architecture, and belongs to the field of image classification, and the method comprises the following steps: preprocessing hyperspectral image data, and dividing the hyperspectral image data into a training set and a test set; constructing a hyperspectral image classification model based on proxy attention and a VIT architecture; training the hyperspectral image classification model by adopting the training set; and inputting the hyperspectral data in the test set into the trained hyperspectral image classification model, reasoning to obtain a prediction category, and evaluating the obtained prediction category through an evaluation index. According to the invention, the agent attention encoder based on agent attention is constructed, hierarchical feature extraction is carried out through the agent attention encoder of L-layer cascade, layer-by-layer fusion and enhancement of space-spectrum information in an initial feature coding matrix are realized, the calculation complexity is significantly reduced while the strong representation capability is maintained, and the method has the advantages of high efficiency, high resolution and the like. And good balance between the calculation efficiency and the feature expression capability is realized.

The invention discloses a hydraulic motor fault diagnosis method and system based on heterogeneous asynchronous data fusion, and the method comprises the steps: collecting heterogeneous asynchronous data of a sensor network, carrying out the preprocessing of the data, inputting the data into a parallel dynamic pruning residual network, extracting features, and carrying out the preliminary fusion, thereby obtaining an initial feature plane; and inputting the initial feature plane into a multi-connection neural network, fusing data, extracting features, obtaining a final feature plane, using the fused feature data as a training set, and building a feature classification model to test and evaluate the performance of the feature classification model. According to the method, the limitation of a traditional fusion model in asynchronous processing of high-frequency vibration signals and low-frequency thermodynamic data of a hydraulic system is effectively overcome, the robustness of key fault features in a strong noise environment is remarkably improved, and rapid virtual-real mapping of bench test simulation data and online monitoring data is realized; typical faults such as plunger pair abrasion and valve plate cavitation of the hydraulic motor can be accurately supported, and safety guarantee is provided for equipment life prediction and safety control.