181 results about "Noise reduction algorithm" patented technology

The invention discloses a partial discharge detection method and device, and the method comprises the steps: outputting a multi-mode signal matrix after noise reduction through employing a self-adaptive noise reduction algorithm according to a multi-mode sensing signal during the operation of a generator; based on the multi-modal signal matrix, a space-time convolutional neural network is adopted to extract discharge features, meanwhile, a topological relation between signal time-frequency features and sensor space is captured, and a multi-modal feature fusion tensor is output; according to the multi-modal feature fusion tensor, a model is generated through a dynamic map, manifold learning and a particle swarm optimization algorithm are combined, and a dynamic fault map containing discharge intensity, phase and frequency point distribution characteristics is output; and based on the dynamic fault map, calculating a real-time discharge danger coefficient by using a risk prediction model, and outputting a discharge grading early warning instruction and a maintenance priority sequence. According to the embodiment of the invention, high-precision and traceable discharge detection and graded early warning can be realized.

The invention relates to the technical field of constructional engineering safety monitoring, and discloses a constructional engineering safety monitoring method and system based on machine learning, and the method comprises the steps: obtaining an original detection data set; according to the original detection data set, adopting a noise reduction algorithm to obtain a monitoring data set; according to the monitoring data set, a clustering algorithm is adopted to divide a building structure into a plurality of areas, and a high-risk area is obtained; according to the high-risk area, time sequence data are extracted, a time sequence analysis method is adopted to calculate a quality index change trend, and a trend analysis result is obtained; according to the trend analysis result, analyzing quality index relevance between different areas through a relevance mining algorithm to obtain a potential chain reaction risk assessment report; and inputting the potential chain reaction risk assessment report into a pre-trained building engineering safety monitoring model to obtain a building structure health trend report. According to the method, self-adaptive building quality safety dynamic global analysis can be realized.

The invention discloses a rescue method and system of a rescue robot for exploration, and relates to the technical field of underground space rescue, and the method comprises the following steps: obtaining multi-path acoustic echo data of a karst cave and motion track data of the robot, and constructing a three-dimensional point cloud model according to the multi-path acoustic echo data and the motion track data of the robot; and identifying unmatched data in the multi-path acoustic echo data and the robot motion trail data, and taking an area where the unmatched data is located as an abnormal area. According to the method, the spatial resolution of signal acquisition is enhanced through the multi-microphone array, robust acoustic fingerprints are extracted in combination with a noise reduction algorithm and short-time Fourier transform, high-confidence human body sound source signals are screened out by using a feature template matching mechanism, accurate extraction and recognition of human body acoustic features in a karst cave complex noise environment are realized, and the accuracy of human body acoustic feature recognition is improved. The problem of misjudgment caused by confusion of sound source features and environmental noise in a traditional method is effectively solved.

The invention discloses a multi-mode perception and reinforcement learning engineering machinery intelligent regulation and control method and system, and belongs to the technical field of intelligent control and industrial automation. According to the system, stratum data within the range of 20-50 meters in front of a shield tunneling machine are collected in real time through multi-mode sensing equipment such as a distributed optical fiber sensor, a cutterhead vibration sensor and an electromagnetic wave radar, and original data are processed by adopting a wavelet-Fourier combined noise reduction algorithm; and inputting the processed multi-modal data into a reinforcement learning intelligent decision-making model based on a CNN-LSTM hybrid network architecture. The model is trained through a dynamic reward function, and weight coefficient combinations can be automatically switched according to different construction scenes. According to the system, parameters such as the rotating speed, the thrust and the grouting amount of a shield cutter head are regulated and controlled in real time through the self-adaptive control module, and the slurry utilization rate is increased through the gradient pulse grouting technology. The precision, efficiency and safety of shield construction are remarkably improved, and meanwhile energy consumption and construction cost are reduced.

The invention provides an unmanned aerial vehicle oblique photogrammetry data intelligent processing method and system, and relates to the technical field of unmanned aerial vehicle photogrammetry, and the method comprises the steps: carrying out the preprocessing of an oblique photogrammetry original image, obtaining the exterior orientation elements of a camera, carrying out the scene semantic segmentation through a deep convolutional neural network, adaptively determining the optimal three-dimensional reconstruction parameters of each region, and obtaining the optimal three-dimensional reconstruction parameters of each region. On the basis of multi-view image matching, feature point parallax information is calculated to generate dense point clouds, different areas are processed by adopting an adaptive grid division strategy and a corresponding noise reduction algorithm, and finally, a three-dimensional scene model with real textures is constructed, so that high-precision three-dimensional reconstruction of different scene areas is realized.

The invention relates to the technical field of earphone noise reduction processing, and discloses a motorcycle riding earphone noise reduction method combining different noise reduction algorithms, comprising the following steps: acquiring an original audio signal, and distributing the original audio signal to each processing module; a preprocessing signal is generated, voice activity information is determined, and environmental noise energy information is calculated; according to the environmental noise energy or the auxiliary information, adaptively adjusting a high-low noise energy threshold value; the original signals are input into the traditional and AI noise reduction module in parallel to generate two paths of noise reduction signals; based on the environment noise, the voice information and the threshold value, performing weighted fusion to generate an output signal; and carrying out equalization and compression processing on the output signal, and driving the loudspeaker to output. According to the invention, comprehensive judgment on environmental noise energy information, voice activity information and riding speed is introduced, an output signal of a traditional noise reduction module is set to be zero in a high-speed voice-free environment, and AI noise reduction is combined, so that tone quality distortion of a traditional noise reduction algorithm in a noise environment is effectively avoided.

The invention relates to the technical field of Internet of Things communication, and provides a low-power-consumption semiconductor communication method based on Internet of Things, which comprises the following steps: acquiring temperature, electromagnetic interference intensity and signal attenuation coefficient data through a multi-mode sensor node, and generating an original environment data set; performing dynamic noise classification on the original environment data set, dividing noise levels and marking noise types; performing hierarchical processing on the data by adopting a multi-stage self-adaptive noise reduction algorithm to generate a preprocessed data set; extracting dynamic characteristic parameters of a communication channel, constructing an energy efficiency prediction model, and dynamically calculating a combination of optimal transmitting frequency, power and modulation order; generating a multi-system modulation signal, superposing a dynamic coding strategy, and adjusting the symbol block length and the error correction code density; and the receiving end analyzes the signal quality index, feeds back the signal quality index to the transmitting end to optimize the parameters and updates the energy efficiency prediction model. According to the invention, the energy efficiency ratio and the anti-interference capability of the Internet of Things communication system can be improved.

The invention relates to the technical field of electric power topology identification, and discloses an electric power topology rapid identification method based on depth map matching, and the method comprises the following steps: S1, carrying out the multi-source heterogeneous data fusion and adaptive noise reduction, and employing an adaptive threshold noise reduction algorithm; s2, modeling a dynamic space-time diagram; s3, attention-driven graph matching is carried out, and a two-channel graph neural network is used; and S4, updating the incremental topology library. For the condition that the real-time data of the low-voltage power distribution network generates significant noise due to equipment aging and acquisition errors, a self-adaptive threshold noise reduction algorithm is adopted, high-frequency noise components are accurately filtered out according to a threshold value, missing data are intelligently repaired in combination with a space-time Kriging interpolation method, the purity and integrity of the data are ensured, and a data basis is provided for subsequent topology recognition.

The invention discloses an electric pump control method and system based on instruction perception, and relates to the technical field of electric pump control, comprising: starting a voice recognition module, a wireless communication module and a sensor acquisition module, and establishing a bidirectional communication link with a remote controller; an AI-driven noise reduction algorithm is adopted to pre-process the environment audio collected by the microphone array, and an improved hidden Markov model is utilized to perform keyword recognition on the pre-processed audio to obtain a voice instruction signal; the remote controller sends a control instruction to a receiving end, analyzes the instruction content, records the current communication channel quality, dynamically adjusts the transmitting power, the frequency hopping strategy or the coding mode according to the channel quality, and ensures stable communication between the remote controller and the electric pump; and the main control unit simultaneously receives a voice command signal and a remote controller command signal, sets a command priority rule in combination with data acquired by the sensor, and executes corresponding operation according to the priority.

An electrocardiosignal anomaly recognition and analysis system based on a multilevel clustering algorithm is characterized by comprising an electrocardiosignal acquisition module, a data transmission module, a data storage management module, a data processing module, an electrocardiosignal denoising module, an electrocardiosignal analysis module and an electrocardiosignal judgment module. According to the electrocardiosignal anomaly recognition and analysis system based on the multi-level clustering algorithm, electrocardiosignal data are collected through the external lead electrodes, noise reduction is conducted on the collected electrocardiosignal data through the reconstruction noise reduction algorithm, then the electrocardiosignal data are clustered through the multi-level clustering algorithm, and the electrocardiosignal anomaly recognition and analysis system based on the multi-level clustering algorithm is obtained. And performing abnormal electrocardiosignal data output based on the standard abnormal electrocardiosignal data.

The invention discloses a submarine cable fault positioning simulation verification system and a calibration method thereof, and belongs to the technical field of submarine cable detection. According to the system, a multi-source sensor array is deployed for data acquisition, and a multi-stage adaptive noise reduction algorithm, a time-frequency domain feature fusion technology and a land-sea environment electromagnetic field mapping model are adopted, so that magnetic field distribution features of a submarine cable in an actual marine environment can be simulated with high fidelity, and accurate positioning of a fault position is realized. By establishing a standardized calibration process, effective migration of simulation data and marine measured data is ensured, and the technical problems of complex marine noise interference and extrapolation distortion of a laboratory environment verification result are effectively solved. The scheme has the remarkable advantages of high positioning precision, strong anti-interference capability, accurate environment mapping and the like.

The invention relates to the technical field of carbon emission source inversion calculation, in particular to a construction site carbon emission source inversion calculation method based on an atmospheric diffusion model. According to the technical scheme, the method comprises the following steps that a spatial diffusion model is constructed, meteorological parameters of a construction site are obtained through meteorological monitoring equipment, pollutant concentration distribution is calculated based on an atmospheric diffusion model, and the model corrects vertical diffusion by introducing a mirror image source item; performing dynamic noise reduction on the data: performing filtering processing on the original monitoring data, and balancing noise suppression and real fluctuation retention through a dynamic adjustment algorithm; performing iterative inversion calculation; and comprehensive verification: combining theoretical simulation with actual detection. Through the dynamic coupling meteorological-emission model, the adaptive noise reduction algorithm and the multi-source verification mechanism, accurate, efficient and low-cost inversion of the carbon emission source of the construction site is realized, and reliable technical support is provided for real-time carbon management, carbon transaction and green construction.

The invention provides a full-spectrum water quality multi-parameter dynamic inversion method and a model construction method thereof, and belongs to the technical field of spectral analysis based on machine learning. By constructing a closed-loop optimization architecture of noise reduction, wavelength optimization, turbidity correction and adaptive modeling, high-precision detection of four water quality parameters of chemical oxygen demand, total organic carbon, total nitrogen and nitrate nitrogen is realized. According to the method, a variational mode decomposition and improved threshold translation invariant wavelet combined noise reduction algorithm is provided, a kurtosis-correlation coefficient dual index is adopted to screen noise components, a multi-translation average strategy is combined to suppress a pseudo-Gibbs phenomenon, and meanwhile, a global parameter collaborative tuning mechanism based on Bayesian optimization is designed; a cross-module hyper-parameter coupling space is established in the whole process of noise reduction, wavelength optimization, turbidity correction and modeling. Through a multi-dimensional feature enhancement mechanism, the model feature representation capability is improved by 35%, and an innovative solution is provided for multi-parameter online detection in a complex water quality scene.

The invention belongs to the technical field of earphone noise reduction, and provides a noise reduction regulation and control method for an earphone and a noise reduction earphone. The method comprises the following steps: acquiring a first environment sound signal of an environment area, and extracting human voice features from the first environment sound signal to construct a human voice feature template library; in response to a noise reduction regulation and control instruction, collecting a second environment sound signal of the environment area, converting the time domain signal into a frequency domain signal through Fourier transform, and extracting a frequency spectrum feature from the frequency domain signal; performing matching calculation on the spectrum features and a human voice feature template library to determine environment voice signals, human voice signals and noise signals; an active noise reduction algorithm is adopted to generate offset sound waves opposite to the noise signals in phase, the human voice signals are amplified, and the offset sound waves and the amplified human voice signals are output through a loudspeaker of the earphone. According to the invention, effective noise filtering and clear transparent transmission of specific human voice are realized, smooth communication can be realized without taking off the earphone, and the use experience is improved.

The invention provides a GIS equipment withstand voltage test system and method, and relates to the technical field of GIS equipment withstand voltage test. Comprising the steps of obtaining parameter information and installation environment data of GIS equipment, constructing a three-dimensional digital twinborn model of the GIS equipment in combination with an electric field simulation engine, generating a withstand voltage test scheme and planning an optimal path. Acquiring partial discharge data and insulation resistance data under different voltage stresses by using an AI noise reduction algorithm according to the optimal path, carrying out multi-modal fusion processing on the partial discharge data and the insulation resistance data, extracting features by using a deep learning model, obtaining insulation defect and performance attenuation data, and carrying out multi-modal fusion processing on the insulation defect and performance attenuation data. And carrying out space mapping association with the three-dimensional digital twinborn model to generate a GIS equipment withstand voltage test evaluation report, thereby providing a powerful basis for equipment operation and maintenance. According to the invention, the three-dimensional digital twinborn model is constructed to generate the GIS equipment voltage withstanding test evaluation report, so that the accuracy of the GIS equipment voltage withstanding test is improved.

The invention discloses an inter-floor noise active suppression system and method.The system comprises a noise monitoring module, a signal processing module, a sound wave emission module and an intelligent control module, and the noise monitoring module, the sound wave emission module and the intelligent control module are all electrically connected with the signal processing module; the noise monitoring module comprises a broadband condenser microphone array, a vibration acceleration sensor and a beam forming processor. Through combination of acoustic monitoring, signal processing and sound wave emission technologies, upstairs noise can be monitored in real time, reverse sound waves with opposite phases and equal amplitudes are generated, so that the noise is effectively counteracted, and through combination of the multi-mode acoustic sensor array and a reverse sound wave noise reduction algorithm, real-time monitoring and quick response to floor noise are realized, and the noise reduction efficiency is improved. The real-time performance and accuracy of the noise reduction effect are ensured, and a quiet living and working environment is created for residents and office staff.

The present invention relates to an edge collaboration system of an intelligent fusion terminal in an electric power substation, characterized in that it includes several intelligent fusion terminals and a cloud. Specifically: a data hierarchical processing mechanism is introduced in the intelligent fusion terminal, and real-time processing is realized by dividing sensor data into high priority and low priority; and based on a noise reduction algorithm for time series anomaly detection, the sensor data is subjected to noise reduction processing; a dynamic collaboration network is established among several intelligent fusion terminals in the substation according to the distribution priority of the computing amount, and a hierarchical task unloading collaboration mechanism is adopted to dynamically allocate computing tasks to other intelligent fusion terminals; the cloud is based on an LSTM model, and the load change of the substation is predicted in real time according to the processed sensor data; a distribution diagnosis model based on a graph neural network is introduced in the cloud; and fault self-healing is realized through a dynamic collaboration network of intelligent fusion terminals. The present invention significantly improves the intelligence level of the substation and the operation efficiency of the power grid.

The invention particularly relates to a voice noise reduction method suitable for different noise environments, and relates to the technical field of voice signal processing. Analyzing a noise environment and characteristics; adaptive selection of a noise reduction algorithm; adaptive noise reduction processing is executed; and noise reduction effect evaluation and iterative optimization are carried out. According to the invention, refined noise classification is matched with the algorithm, so that the limitation of a traditional single algorithm in a complex environment is solved; the steady-state noise is dynamically counteracted by adopting an adaptive filter, the frequency domain gain of the unsteady-state noise is adjusted in real time through Wiener filtering, and the pulse noise is subjected to dynamic threshold processing by combining median filtering and wavelet transform; especially, the design of cooperation factors in wavelet transform can accurately distinguish signal details and noise: intensively suppress pulse noise, loosely retain details such as voices and consonants, and realize dynamic balance of noise reduction intensity and signal distortion.

The invention discloses a terahertz gas insulated switch measuring device and method based on artificial intelligence, the terahertz gas insulated switch measuring device comprises a data imaging module and a data processing module, the data imaging module comprises a terahertz wave emission source, a lock-in amplifier module, a quasi-optical system and a high-sensitivity detector, and the data processing module comprises an upper computer; when the insulation degree of the gas insulated switch is measured, the data imaging module terahertz wave emission source generates terahertz waves, the terahertz waves are focused through the quasi-optical system and transmitted to the gas insulated switch equipment, the high-sensitivity detector receives the terahertz waves passing through the gas insulated switch equipment and converts the terahertz waves into electric signals, and the electric signals are transmitted to the gas insulated switch equipment. Further, the lock-in amplifier module amplifies and filters the electric signals output by the high-sensitivity detector, the data processing module receives the electric signals output by the lock-in amplifier module, defects are recognized through a noise reduction algorithm and a deep learning model, and a data processing result is displayed; the problems that traditional detection is low in precision, small in range, low in efficiency and the like are solved.

The invention discloses an elevator steel belt damage detection and quantitative analysis system based on eddy current and magnetic flux leakage dual-mode detection, and aims to solve the problems of single-mode information loss and serious industrial field strong noise interference in the existing steel belt detection. The system synchronously integrates an eddy current sensor, a magnetic flux leakage sensor and an encoder through a multi-probe array adapter; and multi-dimensional damage physical information in the steel strip is obtained. An improved wavelet packet transform-empirical mode decomposition (WPT-EMD) collaborative noise reduction algorithm is adopted, and in combination with a sub-band energy entropy and a self-attention mechanism, non-stationary mechanical noise is effectively filtered out. A multi-rule feature extraction engine is used for extracting smooth residual errors, derivative mutation and other features, a support vector machine (DE-SVM) model introducing physical priori knowledge weights is constructed, and the small sample recognition problem is solved in combination with a virtual sample generation technology. The method can realize high-precision positioning and quantitative evaluation of steel strip damage under complex working conditions, and has the characteristics of strong anti-interference capability, high identification accuracy and good generalization performance.

The invention discloses a facility agriculture voice interaction system based on a large model, and relates to the technical field of voice interaction. Comprising a voice signal acquisition and enhancement module, a voice recognition and semantic standardization processing module, a feature extraction and emergency degree quantification module, an intelligent evaluation and emergency instruction judgment module and a priority scheduling and response execution module, voice signals are continuously monitored through a multi-microphone array of the wearable terminal, the direction of a voice source is enhanced by using a far-field beam forming technology, the signals are processed through a time domain / frequency domain noise reduction algorithm, and finally voice streams are cached in real time with millisecond-level sampling precision. Through the multi-microphone array and far-field beam forming technology, voice capture is optimized, noise is reduced, real-time voice recognition and semantic analysis of the edge server are combined, the priority of emergency instructions is quantified, it is ensured that resources are rapidly scheduled to process emergency tasks, the agricultural production efficiency is improved, and crop losses are reduced.

The invention discloses a PLC electric control cabinet control method, and belongs to the technical field of electrical control. The method specifically comprises the following steps: S1, multi-dimensional parameter fusion acquisition: acquiring multi-dimensional data through a multi-type sensor, setting a hierarchical acquisition period according to parameter types, storing multi-source data in a classified and layered manner, and performing traceability judgment on abnormal data; s2, parameter preprocessing and feature extraction: different noise reduction algorithms are adopted for different signal types, and depth features of data are extracted through multi-step processing; electrical parameters, physical state parameters, environmental parameters and time sequence label parameters are collected through multiple types of sensors, the equipment operation state and external influence factors are comprehensively captured, compared with traditional single parameter collection, the real working condition of the system can be reflected more accurately, and complete data support is provided for subsequent control and diagnosis. And a hierarchical acquisition period is set according to parameter types, so that resource waste or key data omission caused by one-step acquisition is avoided.

The invention relates to the technical field of time length detection, and discloses a news simulcasting time length detection alarm method based on voice recognition, which comprises the following steps: S1, noise reduction: integrating an advanced digital signal processing technology and a professional noise reduction algorithm to form a refined processing system adaptive to'news simulcasting 'audio, according to the system, aiming at collected useless components such as environment noise and equipment noise, noise characteristics are captured through a multi-dimensional signal analysis model, an effective information core identifier is defined, a differentiated characteristic database of signals and noise is constructed, and a technician sets scientific noise reduction parameters according to the differentiated characteristic database. And accurate noise identification and stripping: integrating an advanced digital signal processing technology and a professional noise reduction algorithm, capturing noise characteristics through a multi-dimensional signal analysis model, accurately stripping environmental noise (such as studio echo and external interference sound) and equipment noise (such as sound recording equipment floor noise and transmission link electromagnetic interference), and simultaneously retaining the original texture of effective audio.

The invention discloses a teaching management resource matching method and system based on deep learning, and the method comprises the steps: obtaining the initial teaching management data through obtaining the text data, video data, audio data and Internet of Things equipment data related to teaching, carrying out the noise reduction of the initial teaching management data through a mixed noise reduction algorithm, and obtaining a noise reduction result; noise reduction teaching management data is obtained; performing feature normalization processing on the noise reduction teaching management data based on wavelet transform and BERT semantic embedding to obtain feature teaching management data; and constructing a CNN-LSTM dual-channel model, and inputting the feature teaching management data into the CNN-LSTM-GAN dual-channel model for identification to obtain a teaching resource matching scheme. Precise matching of teaching resources with student learning requirements and classroom teaching objectives is realized, personalized resource recommendation is supported, and teaching pertinence is improved.

The invention discloses a power cable laying stress monitoring device which comprises a stress sensing module, a signal processing module, a data transmission module and a power supply module. The stress sensing module adopts an integrated FBG (Fiber Bragg Grating) sensor array of a flexible mounting belt, and realizes non-damage coating of a cable with the diameter of 3-200mm through an adjustable fastening device; the signal processing module comprises a low-noise optical fiber amplifier and a self-adaptive noise reduction algorithm, and is used for realizing + / -0.5 mu epsilon strain resolution and 5ms dynamic response; the data transmission module adopts a multi-band 4G / 5G communication module and supports AES-256 encryption and edge calculation; and the power supply module adopts a solar energy-lithium iron battery hybrid power supply system to realize 30-day uninterrupted operation. Through collaborative design of the FBG sensing array and the flexible mounting belt, the device solves the problems of poor mounting adaptability, low precision, response lag and the like of a traditional stress monitoring device, and is particularly suitable for real-time stress monitoring in the power cable laying process.

The invention discloses an end-to-end automatic driving method, and aims to solve the problems that an existing modularized automatic driving system is easy to cause information loss and error accumulation, and an end-to-end method does not fully consider model cognitive deviation and the potential influence of diversified noise on automatic driving in a complex driving scene. According to the method, by constructing an R-Mama module, multi-scale global features are fully extracted, and downstream task performance such as target detection and path planning is improved; a sensing network guided by a self-adaptive structured noise reduction coding-decoding network is designed, noise is randomly added and a noise reduction algorithm is self-adaptively designed for noise reduction according to scene characteristics, and dynamic target detection, tracking and online map construction are realized. According to the method, the anti-interference capability and robustness of the system in a complex environment can be enhanced, the understanding and prediction precision of a dynamic scene is improved, and finally efficient collaborative optimization of sensing, prediction and planning is realized.

The invention discloses an automatic detection method and system for the boundary dimension of a shield segment, and relates to the field of civil engineering, and the method comprises the steps: laying quantum dots on the surface of a to-be-detected shield segment, and exciting the quantum dots through ultraviolet laser to generate a fluorescence spectrum; the method comprises the following steps: receiving an optical signal penetrating through a medium gap in a to-be-detected shield segment through a single-pixel barrel detector to form a reflected light intensity sequence, and inputting the reflected light intensity sequence and a speckle pattern into a correlation reconstruction algorithm to generate a three-dimensional point cloud of a shielding area; and inputting the three-dimensional point cloud and the deformation compensation parameters into a deep learning network, generating a deviation thermodynamic diagram of the boundary dimension of the to-be-detected shield segment, automatically marking an area exceeding a tolerance threshold value based on the deviation thermodynamic diagram, and generating a detection result of the boundary dimension of the to-be-detected shield segment. According to the method, the internal medium gap is reconstructed through the single-pixel barrel detector and the basis tracking noise reduction algorithm, in a bolt hole oil stain shielding scene, compared with a traditional method, the reconstruction porosity error is reduced, and online analysis of the full-surface size error is achieved.

The invention discloses a partial discharge voiceprint multi-mode detection method, which comprises the following steps that voiceprint matrix detection nodes are set, on-site multi-mode signals are collected, and the voiceprint matrix detection nodes comprise an ultrasonic sensor, an infrared thermal imager, a temperature and humidity sensor and a camera; preprocessing the signal, extracting voiceprint features, and constructing a multi-modal feature vector; calculating three-dimensional coordinates of a partial discharge source, and performing spatial mapping verification by combining video imaging; and fusing the features and carrying out classification identification, and outputting a partial discharge type and a risk level. Through voiceprint matrix arrangement, temperature, humidity and video imaging data are combined, and the sensitivity and the anti-interference capability of partial discharge detection are remarkably improved. A noise reduction algorithm and a voiceprint feature fusion technology are adopted, so that complex environmental noise and real discharge signals can be effectively distinguished; the multi-modal data collaborative analysis can evaluate the insulation state of the equipment in real time, the false alarm rate is reduced to below 2%, and unmanned intelligent inspection of the transformer substation is supported.

The invention relates to the technical field of data noise reduction, and discloses a safety alarm noise reduction method and system based on a data weaving technology, and the method comprises the steps: building a unified logic data view based on a data weaving architecture; deploying an AI-driven data intelligent agent component system, automatically identifying and adapting a changing API and a new data source, and fusing metadata, threat intelligence and infrastructure operation state information; constructing an attack chain panorama and a behavior causal chain model; a semantic noise reduction algorithm, time sequence clustering analysis and a confidence scoring mechanism are adopted to perform de-duplication alarm, false alarm identification and low-risk event filtering on the security alarm; and alarm de-duplication and priority adjustment are carried out to generate a linkage processing strategy. According to the invention, the problems of alarm flooding, high false alarm rate, low processing efficiency and the like in the existing security alarm system are solved.

The invention relates to the technical field of medical image processing, in particular to a dynamic ultrasonic noise reduction model training method and a dynamic ultrasonic noise reduction method and device. The method comprises the following steps: acquiring an ultrasonic video covering multiple clinical scenes such as antral pneumatosis, hydrops, solid contents, long antral antrum and cardiac anatomical occlusion; marking the precise boundary of the organ frame by frame by a clinical expert, segmenting and complementing a missing section, optimizing the contour of a fuzzy region, and inhibiting or erasing artifacts; a deep learning model is trained based on the labeled data, so that the model can call an exclusive noise reduction algorithm according to organ categories (the antral sinuses adopt morphological reconstruction filtering guided by ellipse / long axis prior, and the heart adopts time-space joint deblurring driven by a cardiac cycle and combining motion compensation and anisotropic filtering); and weighted noise reduction is carried out on the suspected lesion or the user indication area. In the reasoning stage, a high-fidelity and low-noise video stream can be output in real time by inputting any dynamic ultrasonic video, and organ boundary metadata is attached.