981results about How to "Suppress interference" patented technology

The invention discloses a remote sensing image change detection method based on an adaptive Transform and a deformable convolution. The method comprises the following steps: firstly, establishing a twin feature extraction encoder fusing adaptive Transform and deformable convolution for extracting multi-scale local-global features of a dual-temporal remote sensing image; secondly, an adaptive Cross-Transform high-order feature interaction module is designed, and the adaptive Cross-Transform high-order feature interaction module is used for extracting high-level semantic change information and inhibiting irrelevant information interference; thirdly, providing a difference extraction module to generate different levels of difference feature maps so as to obtain a multi-scale change region; and finally, constructing a multi-scale fusion decoder, fusing the multi-level feature maps layer by layer, and further enhancing the feature expression of the change region. According to the method, the advantages of the Transform and the CNN are combined, the designed adaptive Transform and adaptive Cross-Transform can effectively extract the global features of the change region, and the adopted deformable convolution can effectively extract the local features of the change region, so that the overall performance of change detection is improved.

The invention discloses a PPY-YOLO-based steel surface defect detection method and a PPY-YOLO-based steel surface defect detection system. The PPY-YOLO-based steel surface defect detection method comprises the following steps: step 1, preprocessing a steel surface defect image; 2, a neck structure Faster-Neck, a deep learning module C3K2Faster and a spatial pyramid pooling module SPPFSim are introduced into the YOLOv11 network, and a PPY-YOLO target detection model is obtained; step 3, training a PPY-YOLO target detection model; and 4, acquiring a steel image, and inputting the steel image into the trained PPY-YOLO target detection model to obtain the category and positioning information of the steel surface defects. A new model is constructed, the backbone network and the neck network are improved, the backbone network is formed by connecting the input end of the C2PSA module with the output end of the SPPFSim module, and an up-sampling group and a down-sampling group are added to the neck network, so that the model has a relatively good detection effect.

The invention discloses a typical ground object target segmentation method based on multi-source data attention feature fusion, and the method comprises the following steps: (1), obtaining a data set containing a high-resolution optical image, an SAR image and a corresponding semantic tag image, and carrying out the preprocessing of the data set; (2) constructing a typical ground object target segmentation model based on multi-source data attention feature fusion, and realizing semantic segmentation of a typical ground object target; and (3) taking the preprocessed data set obtained in the step (1) as the input of the model constructed in the step (2), training the model based on a weighted cross entropy loss function, and realizing typical ground object target segmentation based on multi-source data. The invention provides an adaptive fusion algorithm based on an attention mechanism by comprehensively considering the observation advantages of SAR and optical data for different ground object targets, and better combines the respective advantage characteristics of an optical image and an SAR image, thereby achieving a better semantic segmentation effect in the field of ground object classification.

The invention provides a railway signal monitoring data generation method and system, and relates to the technical field of railways, and the method comprises the steps: carrying out the time-space correlation fusion processing of obtained carrier frequency characteristics, response coding information, the receiving time of transponder signal data, train operation speed data and train positioning data, and generating a monitoring data record under a unified coordinate system; adopting a rule-based data compression algorithm to adaptively select a compression rule and a compression rate to perform data compression on the monitoring data record; packaging the compressed monitoring data record into a monitoring data file in a standard format; and the data management server decompresses and decodes the monitoring data file, cleans, classifies and indexes the data in the monitoring data file by adopting a MapReduce parallel computing framework, stores the structured monitoring data into a distributed database, and performs fragmentation and redundant backup management on the database.

The invention provides an overhead transmission line defect detection method based on multi-scale feature fusion, and the method comprises the steps: carrying out the correction and registration of a lead image and a laser point cloud image, carrying out the data enhancement, and constructing a detection data set; the method comprises the following steps: respectively inputting two images into a target detection network for training, adding an attention mechanism in a feature extraction network Backbone so as to carry out correlation calculation on a target, and meanwhile, adopting a multi-scale feature fusion detection method in a backbone part of a YOLOV8 operation network, so that the improvement of precision and accuracy is realized on the premise of ensuring the size of a model, and the accuracy of the model is improved. And the requirement of actual overhead transmission conductor defect detection is met.

The invention provides a sea wave magnetic interference suppression method and device for ocean electromagnetic detection, and relates to the technical field of electromagnetic signal processing, and the method comprises the steps: synchronously collecting original electromagnetic signals through a plurality of annularly arranged magnetic sensors, generating interference parameters through employing the calculated interference signal amplitude and phase information, and carrying out the subsequent interference suppression processing. The method comprises the following steps of: calculating a covariance matrix, performing principal component analysis, extracting spatial characteristics of low-frequency magnetic interference, projecting a signal to an orthogonal subspace outside an interference subspace, removing a low-frequency magnetic interference component, retaining a target signal, performing Fourier transform to convert the signal to a frequency domain, analyzing a frequency spectrum by using a band elimination filter, and setting an interference threshold value. According to the identified interference frequency, an interference suppression signal is generated, the interference suppression signal is returned to the time domain through inverse Fourier transform and is superposed with the target signal to generate a standard electromagnetic signal, the magnetic interference in the ocean detection area can be effectively suppressed, and accurate acquisition and analysis of the standard electromagnetic signal are ensured.

The intelligent factory data acquisition platform comprises a data sensing layer, a multi-protocol transmission layer, a data fusion layer and an application service layer, the data sensing layer is used for collecting equipment state and environment parameter multi-source heterogeneous data through various sensors, preprocessing the collected data and transmitting the preprocessed data to the multi-protocol transmission layer; the multi-protocol transmission layer is used for receiving the preprocessed data transmitted by the data sensing layer, packaging the data through a multi-protocol data transmission unit, then carrying out multi-protocol transmission, and outputting the transmitted data to the data fusion layer; the data fusion layer is used for receiving the multi-source heterogeneous data transmitted by the multi-protocol transmission layer, performing feature extraction and fusion on the heterogeneous data by adopting a self-correcting convolutional recurrent neural network, and outputting fused feature representation to the application service layer; and the application service layer is used for receiving the fusion features output by the data fusion layer, and performing equipment health diagnosis, process parameter optimization and energy consumption statistical analysis intelligent application services based on the fusion features.

The invention relates to the technical field of filters, in particular to an internal shielding component of a power filter, which comprises a shell, a cover plate is mounted at the top of the shell through a cross recessed countersunk head screw, and a metal partition plate is arranged on one side of the inner wall of the shell; the wave-proof sleeve is welded on one side wall of the shell, and the material of the wave-proof sleeve is light nickel-plated copper fabric; when the power supply filter is used, the problem that crosstalk coupling is formed between interference signals can be prevented, the electromagnetic compatibility in the signal transmission process between systems is solved, the anti-wave sleeve is adopted for shielding power supply input of the power supply filter, interference of an input cable on other signals is greatly restrained, and the service life of the power supply filter is prolonged. The electromagnetic compatibility problem of the whole system caused by the fact that high-frequency noise is coupled to a power line is prevented, the shielding effectiveness is improved, the problem of filtering post-stage electromagnetic interference signals is greatly restrained, meanwhile, the heat dissipation area is increased, and the heat dissipation efficiency and effect are improved.

The invention discloses a point switch part detection method and device based on a YOLOv8 model and a medium, and the method comprises the steps: marking a target part in a point switch part image, and generating a target image; the method comprises the following steps: replacing an original backbone network darknet53 of a YOLOv8 model with a Swin transfer network, and introducing a residual attention module into the Swin transfer network; the method comprises the following steps of: constructing an improved YOLOv8 model by combining a spatial pyramid pooling layer on the basis of a Swin transformer network and replacing an original loss function CIoU of the YOLOv8 model with a new loss function Focal-EIoU at the same time; and inputting the target image into the improved YOLOv8 model, and generating a prediction result. And accurate detection of the target part is realized based on the detection result output by the improved YOLOv8 model.

The invention provides a multichannel sound and vibration information fusion speed reducer squeal traceability method, which comprises the following steps of: arranging a plurality of vibration sensors and microphones around a speed reducer, and synchronously acquiring vibration signals and sound signals; converting the multi-channel acoustic vibration signals from a time domain to a frequency domain by adopting short-time Fourier transform and averaging the multi-channel acoustic vibration signals along a time dimension; setting an outburst ratio threshold value to realize howling frequency band screening of the sound spectrum of each channel; summing the screened sound frequency spectrum and vibration frequency spectrum along the channel dimension and carrying out amplitude normalization processing; the processed sound frequency spectrum and the vibration frequency spectrum are multiplied according to corresponding frequency points, and a fusion frequency spectrum of sound and vibration information fusion is obtained; and converting the frequency domain into an order domain to obtain a howling traceability result of the speed reducer. According to the method, after the outburst ratio threshold value is set, professional technicians do not need to analyze data, mutual verification of sound and vibration signals can be achieved, interference of external noise is restrained, fusion of sound and vibration information at the feature level is adopted, and accurate howling traceability can be achieved.

The invention belongs to the technical field of microwave passive devices, and discloses a millimeter wave duplex power divider based on a multilayer metal gap waveguide and a working method thereof, and the millimeter wave duplex power divider comprises a common channel, a low-frequency filtering power divider and a high-frequency filtering power divider; the input port of the common channel is used for being connected with an electromagnetic wave signal source, and the output end of the common channel comprises an upper output channel and a lower output channel. Wherein the first output channel is connected with the input port of the low-frequency filtering power divider, and the second output channel is connected with the input port of the high-frequency filtering power divider; the low-frequency filtering power divider and the high-frequency filtering power divider are stacked up and down and adopt double-layer metal gap waveguide structures; wherein a fourth-order resonant cavity is arranged in the double-layer metal gap waveguide structure so as to realize a duplex power division function; functions of a duplexer and a power divider are integrated, loss caused by machining and assembling errors is greatly reduced, and low insertion loss and high isolation between channels of the duplex power divider are achieved.

The invention provides a sewer internal defect detection method and system based on YOLOv11 improvement, and the method achieves the dynamic weight fusion of multi-level features through a self-adaptive feature fusion module, and improves the detection capability of a model for small targets and complex backgrounds. The selectivity of channel characteristics is improved through an SENet attention mechanism, and background interference is reduced; and introducing an MPDIOU loss function to improve bounding box regression loss. The performance of the improved model in detection tasks of sewer defects such as cracks, obstacles, corrosion, deformation and the like is remarkably superior to that of original YOLOv11, and particularly key indexes such as mAP50 (average precision), Box precision (P), recall rate (R) and the like are remarkably improved. According to the invention, efficient and accurate detection of multiple types of defects in the sewer can be realized, and a better solution is provided for maintenance and management of the sewer.

The invention provides an aerial image target detection method and system based on an attention mechanism and a related device, and the method comprises the steps: inputting an aerial image obtained by an unmanned plane into an aerial image target detection network model, and obtaining a target detection result of the aerial image; wherein a backbone network in the aerial image target detection network model is a YOLOv7 backbone network containing a background suppression module BSAM containing an attention mechanism, a neck structure is a bidirectional weighted fusion pyramid module TWFP, and a detection head structure is a combination of a generalization efficient layer aggregation network GELAN and a visual Swin Transformer model.According to the detection method, the detection process is simplified, the detection speed is increased, and the detection efficiency is improved. And meanwhile, relatively high detection precision is kept.

The invention relates to the field of gas leakage detection, and discloses a gridding directional laser methane detection system, which comprises the following steps of: acquiring area information which comprises a methane absorption signal, environmental data, background gas, an area image and historical leakage data; generating first two-dimensional structure data from the collected regional information, wherein the first two-dimensional structure data comprises a data matrix and a relation matrix; first one-dimensional structure data is obtained based on sorting of the first two-dimensional structure data, wherein the first one-dimensional structure data comprises n data items sorted according to time; and inputting the first two-dimensional structure data into the detection model, and outputting a methane leakage processing strategy. According to the invention, detection and treatment strategies of methane in fuel gas can be realized, and a distribution strategy of the methane laser detection device can be further given.

The invention relates to an Internet of Vehicles multi-agent cooperative sensing method and system. The method comprises the following steps: constructing a collaborative sensing network, wherein the collaborative sensing network comprises a self-agent and a plurality of collaborative agents; acquiring and processing sensing data through the collaborative sensing network; performing feature extraction to obtain intermediate features; self-adaptive sparsification is carried out to obtain sparse features, and the sparse features are compressed and transmitted to a self-agent; performing time sequence feature enhancement on the features of all the agents at the self-agent end; fusing the features to obtain fused features; and constructing an Internet of Vehicles perception model, and realizing perception by the detection model according to the fused features. According to the method, the calculation complexity of traditional global attention is reduced from the square level to the linear level through an adaptive sparsification mechanism, the calculation overhead is remarkably reduced while the multi-agent feature interaction precision is kept, and the method is more suitable for real-time operation on the vehicle-mounted edge equipment with limited resources.

The invention relates to the technical field of signal processing, in particular to a sound source localization method and system based on improved adaptive beam forming. A microphone array is used for collecting sound source signals, and a signal receiving model is established; designing an improved adaptive filtering beam forming algorithm, and optimizing a cost function; solving an optimal weighting vector through a minimization cost function, and introducing an adaptive updating mechanism to dynamically adjust the weighting vector; selecting a reference microphone, calculating time difference of arrival and distance difference of other microphones, and establishing a distance difference equation from the sound source to the receiver; and solving a sound source coordinate and calculating the distance between the center of the microphone array and the sound source. According to the invention, through innovative algorithm design and a self-adaptive mechanism, the precision and stability of sound source localization are significantly improved, and especially, the method shows strong anti-interference capability and adaptability in a complex noise environment. The method can effectively improve the equipment maintenance efficiency in the substation equipment noise monitoring, reduces the fault risk, and provides reliable guarantee for the safe operation of a power grid.

The invention discloses a multi-layer screen cover plate glass interlayer foreign matter detection device and detection method, and belongs to the field of image processing, and the detection method comprises the following steps: image acquisition; extracting a region of interest; feature enhancement; carrying out anisotropic Gaussian filtering; edge detection; foreign matter screening. According to the method, the ROI image is enhanced by applying frequency tuning saliency detection FT and Haar wavelet transform, and an attention mechanism is introduced to fuse features, so that the saliency of tiny and low-contrast defects is enhanced; the influence of noise is reduced by adopting anisotropic Gaussian filtering, and important edge information is reserved and interference is suppressed by adaptively adjusting parameters of a filter.

The invention relates to the technical field of target recognition, in particular to a target detection method and system in a complex environment and a medium. The method comprises the following steps: acquiring an unmanned aerial vehicle aerial image of a target area; and inputting the aerial image into a trained target detection model, and detecting an area in which a target to be detected exists in the aerial image, the target detection model being based on an improved YOLOV8 structure. The method is used for acquiring spatial position information of targets based on coordinate attention capture and cross-level fusion of multi-scale features and based on attention so as to detect dense small targets. According to the method, the target detection capability of dense small targets in a complex environment can be improved.

The invention discloses a signal anti-interference method for a satellite remote sensing application system, and the method comprises the steps: receiving an electromagnetic signal from a remote sensing satellite through a satellite antenna, and carrying out the preprocessing; interference detection is carried out on the preprocessed signals, and when energy of partial scales is abnormally high, it is judged that interference exists in the digital signals; combining features extracted in interference detection into a feature vector, inputting the feature vector into a classification model constructed based on a convolutional neural network for interference classification, outputting the probability of each interference type, and selecting the type with the maximum probability as a judgment result; according to the output judgment result, interference signals existing in the digital signals are suppressed by adopting a corresponding self-adaptive filtering algorithm, and purified signals are obtained; performing inverse transformation on the purified signal by using a mother wavelet function, and reconstructing the signal; and storing the reconstructed signal in a digital file form, and transmitting the reconstructed signal to a ground station or a data processing center for subsequent analysis and application.

The invention discloses an environment-friendly safe low-smoke halogen-free cable, and relates to the technical field of cables, the environment-friendly safe low-smoke halogen-free cable comprises a cable assembly, the cable assembly comprises an outer sheath, the outer surface of the outer sheath is fixedly communicated with an air outlet assembly, a plurality of conductors are arranged in the outer sheath, each conductor is hollow, and the outer surface of the outer sheath is fixedly communicated with the air outlet assembly. A plurality of filling pipes are arranged among the exteriors of the plurality of conductors, and the plurality of filling pipes are hollow; according to the environment-friendly safe low-smoke halogen-free cable, when the temperature of air around the cable is low, cold air enters the hollow conductor from the air inlet, air convection is formed in the conductor, heat is taken away and then discharged from the air outlet, and when excessive heat is generated in the use process of the conductor, the heat can enter the filling pipe through the multiple drainage holes, so that the cable is protected. And heat can flow in the conductor, so that excessive local heat dissipation of the conductor can be prevented.

The invention provides a novel YOLOv10-based safety belt detection method and system, and the method comprises the steps: obtaining image data, and marking the image data, namely marking the position / wearing state of a safety belt; constructing a detection model based on a novel YOLOv10 network, namely replacing partial standard convolution Conv of the Backbone network with local attention regularization convolution; using a separable convolution star module to replace an original feature extraction module of the Backbone network; a CARAFE module is used for replacing up-sampling operation in a Neck network, multi-scale space attention dilation convolution fusion modules are additionally arranged before three sets of feature maps output by the Neck network, and then the feature maps are output to a Head network; and training a detection model by using the labeled image data to obtain a safety belt detection model. When the novel YOLOv10 network is used for safety belt detection, the detection efficiency and performance can be remarkably improved.

The invention relates to the technical field of voltage control, and particularly discloses a voltage control method for a high-reliability high-voltage power supply, and the method comprises the steps: carrying out the input filtering and power conversion boosting processing of a DC voltage signal provided by a power supply, so as to obtain a high-voltage AC signal; a signal processing technology based on deep learning is adopted to perform CEEMDAN decomposition and intrinsic mode feature extraction on a sampled voltage signal, feature saliency modulation dynamic aggregation is performed on each intrinsic mode after decomposition of the sampled voltage signal to suppress noise and interference in the signal, and a pure sampled voltage signal is obtained. And error calculation and MOS tube driving control are carried out in combination with an external control voltage signal, so that accurate adjustment of the output voltage of the high-voltage power supply is realized. Therefore, the precision of voltage control can be effectively improved, and the adaptability to complex and changeable load conditions and interference environments is enhanced, so that the stability and reliability of a high-voltage power supply system are remarkably improved.

The invention discloses a power transmission line icing thickness detection method based on a deep convolutional neural network, and the method specifically comprises the steps: carrying out the image data collection of a power transmission line, and carrying out the manual marking of an image; preprocessing the manually labeled image to generate a data set, and dividing the data set into a training set, a verification set and a test set according to a proportion; selecting a deep convolutional neural network model; training the deep convolutional neural network model by using a training set, evaluating by using a verification set, and testing by using a test set to obtain a power transmission line icing thickness detection model; and deploying the power transmission line icing thickness detection model to a cloud server or an edge computing device, and carrying out icing thickness detection on the power transmission line. According to the transmission line icing thickness detection method based on the deep convolutional neural network, the problems of low detection efficiency, poor accuracy and high cost in the prior art are solved.

The invention provides a ginseng planting area remote sensing recognition method and system based on a spectral index, and relates to the technical field of remote sensing image processing and agricultural resource monitoring. Remote sensing images are preprocessed, spectral information of ginseng sheds and other ground features is extracted and analyzed, and the types of the ginseng sheds are divided into yellow sheds and blue sheds. Based on spectral characteristics and separability analysis, shed identification indexes are constructed, spectral characteristic differences between yellow sheds and blue sheds and other ground objects are amplified through effective wave band combination and calculation, and background interference is inhibited. Through random sample generation and artificial discrimination, precision iteration is carried out to determine an optimal segmentation threshold, and rapid and efficient automatic identification of different types of parameter sheds is realized. The method can effectively identify yellow and blue sheds at the same time, has good regional applicability, effectively inhibits interference of surface features such as bare soil and blue roofs, and has high identification precision and stability.

The invention discloses a sea surface weak target detection method and system based on sea clutter suppression and a storage medium, relates to the field of sea surface target detection, and aims to solve the problems that an existing sea surface target detection method is difficult to effectively suppress sea clutters and is difficult to recognize sea surface weak targets under complex sea conditions. Comprising the following steps: step 1, carrying out wavelet transform on a target echo signal of a radar, and changing the resolution of time and frequency through scale scaling to obtain time-frequency domain components under different scales; step 2, estimating clutters of a distance unit to be detected by using the clutters of the adjacent distance units according to the correlation between the clutters of the adjacent distance units of the sea surface and between multiple observation values of echoes of the same distance resolution unit; step 3, constructing a Hankel matrix for the echo signal, performing singular value decomposition, and eliminating a large singular value corresponding to the sea clutter according to the characteristic that the sea clutter energy in the same Chirp signal is greater than the signal energy to obtain a new Hankel matrix; and 4, echo signal reconstruction is carried out, and a sea surface target is detected.

The invention discloses an unmanned aerial vehicle low-altitude visual angle target detection method and device based on deep learning, and belongs to the technical field of computer vision. The method comprises the following steps: performing data enhancement on a collected target detection object image data set, and dividing the enhanced target detection object image data set into a training set, a verification set and a test set; an attention mechanism module is added in a backbone network of the YOLOV7-OBB model; the YOLOV7-OBB model is trained based on the training set, the YOLOV7-OBB model in the training process is evaluated based on the verification set, and the trained YOLOV7-OBB model is obtained; and inputting the test set into the trained YOLOV7-OBB model, and carrying out target detection to obtain a detection result. By enhancing a target detection object image data set and adding an attention mechanism module, the YOLOV7-OBB model fully learns key feature information of a target, and the detection precision and robustness are remarkably improved.

The invention discloses a charged particle identification method, a charged particle identification system based on a screen grid ionization chamber, computer equipment and a storage medium, and belongs to the technical field of radiation detection. The charged particle identification method comprises the following steps: acquiring signals generated by charged particles at a cathode and an anode of a detector and generating characteristic quantities; constructing particle identification parameters by using the obtained characteristic quantity; generating a two-dimensional spectrogram according to the identification parameters and observing distribution characteristics of different particle events; and identifying different particles according to the distribution characteristics of different particle events in the two-dimensional spectrogram. Particle energy is associated with an ionization range during identification, the problem of limitation of a traditional particle identification technology in a high event rate and complex particle product environment is solved, especially the capability of processing a tritium nuclear event superposition phenomenon in a resonance energy region is improved, and therefore the problem that due to the fact that the energy resolution is low, the probability of tritium nuclear event superposition is low can be solved. And the defect problem in the prior art that the detection range is limited due to the fact that complex particle types cannot be accurately identified is solved.

The invention discloses a TBM cutterhead rub-impact fault diagnosis method based on vibration frequency micro fluctuation. The method comprises the following steps: collecting cutterhead spindle vibration signal data; obtaining each frequency component of the vibration signal according to the frequency spectrum component of the vibration signal; carrying out time-frequency analysis on the component signal by adopting a generalized synchronous compression transformation time-frequency analysis method to obtain a time-frequency distribution result of the component signal; according to a generalized synchronous compression transformation result of each component, extracting an instantaneous frequency ridge line of time-frequency distribution, removing a direct current component of the instantaneous frequency ridge line, calculating a frequency spectrum of the instantaneous frequency, and obtaining an oscillation frequency of the instantaneous frequency; judging whether periodic oscillation exists in the instantaneous frequency according to the frequency spectrum of each instantaneous frequency ridge line; if yes, judging whether the oscillation frequency is equal to the harmonic frequency of the vibration signal; and if the conditions are met, judging that the TBM cutterhead system has a rub-impact fault, and judging the rub-impact condition of the cutterhead according to the oscillation frequency. According to the invention, the accuracy and sensitivity of diagnosis are improved.

The invention discloses an environment sound classification method and system based on metric learning, and the method comprises the following steps: collecting an environment sound audio, and converting the environment sound audio into a Mel spectrogram; performing data enhancement on the Mel spectrogram; an SPA module and an FPL module are constructed, the SPA module or the FPL module is used to process the environment sound audio after data enhancement, a pre-trained convolutional neural network model is used to carry out feature extraction on the processed environment sound audio, and sound features are obtained; inputting the extracted sound features into a representation space; sound features are optimized in combination with comparison, metric learning and cross entropy loss in the representation space, and the optimized sound features are classified; according to the method, the intra-class variation is considered by using metric learning, so that the features of the same kind of samples are closer and the features of different kinds of samples are more separated in the process of carrying out the classification task, and the classification accuracy is effectively improved.

A wind turbine generator variable pitch bearing fault diagnosis method based on signal coding specifically comprises the steps that original vibration signal data of a variable pitch bearing are acquired through a wind turbine generator state monitoring system arranged in a wind power plant; bearing vibration signals are respectively coded into a time-frequency diagram and a position matrix diagram by adopting continuous wavelet transformation and position matrix transformation, and the characteristic information of the signals is enhanced by increasing the visual dimension of data; a global feature extraction module and a local feature extraction module are respectively designed based on a global attention mechanism and a window attention mechanism, so that the model can fully learn feature information of different levels; a dual-channel parallel fusion network model is constructed according to global and local feature extraction modules to fuse global and local feature information, and the purpose of suppressing strong random noise interference is achieved. According to the method, the vibration signal characteristics of the variable pitch bearing in different states can be fully captured, and the problems of poor diagnosis performance and low precision of the variable pitch bearing caused by strong random noise interference can be effectively solved.