71 results about "Spectral subtraction" patented technology

The invention relates to the technical field of voice recognition, in particular to a multi-channel voice conversion and synchronous transmission system applied to simultaneous interpretation. Multi-language voice signals are collected through a multi-microphone array, background noise is dynamically eliminated by adopting a noise suppression technology combining spectral subtraction and deep learning, and a dynamic information source verification symbol is generated to ensure data synchronization integrity; a blind source separation technology is combined with time-frequency analysis and time delay estimation to realize multi-language signal separation and synchronization, and a time sequence is dynamically adjusted through voice activity detection; an end-to-end ASR-NMT-TTS model is constructed to realize voice real-time translation and synthesis, and low-delay transmission is carried out based on a 5G network; the real-time monitoring module is adopted to dynamically adjust the output delay and the signal-to-noise ratio, and the translation delay and the synchronization precision are optimized in combination with user feedback. According to the method, dynamic noise reduction, multi-source synchronization and a self-adaptive feedback mechanism are integrated, the problems of distortion and delay of multi-language simultaneous transmission in a complex noise environment are solved, and the obvious technical synergistic effect and practicability are achieved.

The invention discloses a bimodal neural network and spectral subtraction combined underwater acoustic signal noise suppression method. Comprising the following steps: firstly, constructing a simulation underwater sound data set containing self-noise; thirdly, constructing a bimodal neural network, then training the bimodal neural network until training is completed, and obtaining a platform self-noise estimation model; inputting a to-be-processed noise-containing underwater acoustic signal into the platform self-noise estimation model, and outputting estimation platform self-noise by the model; and finally, based on the self noise of the estimation platform, noise reduction processing is carried out on the noise-containing underwater acoustic signal to be processed by using the improved spectral subtraction method, and then a noise-reduced underwater acoustic signal is obtained. The method combines an advanced deep learning method and a classical spectral subtraction algorithm, has the capabilities of autonomous learning and efficient noise reduction, remarkably improves the signal-to-noise ratio of the underwater acoustic signal, is suitable for noise reduction of the underwater acoustic signal in a non-stationary and complex marine environment, remarkably improves the low-frequency noise suppression capability and the target signal fidelity, and has a good application prospect. The method is suitable for underwater acoustic target detection, underwater acoustic communication and other scenes.

The invention relates to a pipeline leakage detection method based on acoustic signal spectrum noise reduction, and belongs to the technical field of pipeline leakage monitoring. The method comprises the steps that pure background noise signals are collected through an acoustic sensor to serve as noise analysis samples, noise power spectrum estimation is conducted, and average noise power spectrum estimation is obtained; collecting an original acoustic signal containing potential leakage; performing spectrum analysis and phase reservation on the original acoustic signal after framing processing; a continuous noise reduction acoustic signal is output through spectral subtraction operation and signal reconstruction; performing feature extraction on the continuous noise reduction acoustic signal, performing anomaly detection on a current signal segment through a pipeline leakage detection model, and outputting a leakage signal segment; and leakage point positioning is carried out according to the time difference of the leakage signal segment reaching different sensors. Efficient and accurate detection of pipeline leakage is realized, and interference of background noise on a detection result is effectively overcome.

A sensorless method for machine tool chatter detection. A motor torque signal is analyzed in the time domain to determine whether a bit is currently cutting workpiece material. When not cutting material, an air-cut reference signal is stored for later use. When cutting material, the motor torque signal is converted to the frequency domain and filtered in a multi-step process. After removal of the air-cut reference signal via spectral subtraction, and removal of spindle harmonic components, additional filtering is performed to address aliasing and encoder error effects. The aliasing filtering removes artificial peaks in the frequency response spectrum resulting from interaction between sampling frequency and cutting frequency. The encoder error filtering removes frequency response peaks related to encoder design and interaction with motor speed. After filtering, indicator criteria are evaluated to detect chatter, and corrective action is taken when chatter is detected.

To provide a sensorless method for machine tool chatter detection.SOLUTION: A motor torque signal is analyzed in the time domain to determine whether a bit is currently cutting workpiece material. When not cutting material, an air-cut reference signal is stored for later use. When cutting material, the motor torque signal is converted to the frequency domain and filtered in a two-step process. The first filtering step includes removal of the air-cut reference signal via spectral subtraction, which removes noise and harmonics not directly related to chatter. The second filtering step includes further removal of spindle harmonic components at frequencies known from the spindle speed provided by the machine controller, where this step removes harmonic contents without distorting nearby contents. After filtering, two indicator criteria are evaluated to detect chatter, including a magnitude of the filtered torque signal and a magnitude ratio of the filtered torque signal to the air-cut reference signal.SELECTED DRAWING: Figure 3

The invention discloses a two-stage voice noise reduction method and device based on enhanced spectral subtraction and Kalman filtering, and a storage medium. The method comprises the following steps: performing preliminary noise reduction on a voice signal in a frequency domain through enhanced spectral subtraction; speech signal state space model parameters are identified based on a linear predictive coding algorithm, and secondary noise reduction is performed on the sound signals in a time domain by using a Kalman filtering algorithm; and repeatedly carrying out multiple parameter identification and secondary noise reduction processes on the frame-by-frame signals, and then outputting noise-reduced voice. According to the method, the environmental noise can be effectively filtered in a complex noise environment, and the voice signal quality is improved. The method is suitable for single-microphone noise reduction of the artificial cochlea, efficient noise reduction can be achieved in the artificial cochlea with limited computing resources, and the sound perception ability of a wearer is improved.

The invention relates to the technical field of speech enhancement, in particular to an audio optimization method and system applied to speech recognition, and the method comprises the steps: collecting audio data in real time, and uniformly dividing the audio data into audio frames; for each audio frame and each neighbor frame of the preset audio frame, evaluating whether a noise spectrum is re-selected when the audio data is enhanced by adopting a spectral subtraction method; if reselection is carried out, each mode of the audio frame is acquired; selecting a noise mode from the modes of the audio frame, and obtaining the number of lagging frames of other modes; obtaining a frequency spectrum change characteristic value and a voice characteristic value of each mode, and selecting a main voice mode; acquiring audio mode characteristic values of other modes except the main voice mode; and a noise frame and a new noise spectrum are obtained. The invention aims to improve the accuracy of voice recognition by improving the accuracy of noise frame selection and enhancing the audio enhancement effect of the spectral subtraction for the voice features of the audio frames.

The invention relates to the technical field of voice communication, in particular to a voice signal processing method and device, equipment, a storage medium and a computer program product. The method comprises the following steps: performing real-time noise classification on a received original voice signal based on a preset deep neural network noise recognition model; according to the noise classification result, selecting a corresponding combined noise reduction model to perform multi-stage noise reduction processing on the original voice signal to obtain a noise-reduced voice signal, the combined noise reduction model comprising one or more of a deep neural network-based noise reduction module, a Wiener filtering module and a spectral subtraction module; and on the basis of time domain feature extraction and frequency domain feature extraction, voice feature enhancement processing is performed on the noise-reduced voice signal to obtain the target voice signal, so that the transmission quality of the voice signal is improved.

The invention discloses a digital stethoscope system based on a piezoelectric film, and relates to the technical field of medical electronics and biomedical engineering. An existing digital stethoscope based on a piezoelectric film has the challenges of signal noise and interference, limitation of a signal processing algorithm, nonlinear response and time migration, insufficient diagnosis accuracy and automation and the like. A piezoelectric film and microphone dual audio acquisition mode is fused, various physiological signals such as low-frequency vibration and high-quality heart and lung sound are captured, and background noise is removed through a self-adaptive filter and a frequency spectrum subtraction technology; an MFCC feature extraction and dynamic time warping algorithm is utilized to carry out fine feature analysis and time axis alignment on the preprocessed audio signal, and high-precision matching with a storage template is realized; different physiological states are accurately distinguished through intelligent feature processing, real-time and visual diagnosis feedback is achieved through wireless transmission and state indication, and the reliability and efficiency of medical detection are improved.

A sensorless method for machine tool chatter detection. When the machine tool spindle is running, a spindle motor torque signal is analyzed in the time domain to determine whether a bit is currently cutting a workpiece. When not cutting, an air-cut reference signal is stored for later use. When cutting, the spindle motor torque signal, along with positioning servo motor signals, are converted to the frequency domain and filtered. Filtering steps include removal of the air-cut reference signal via spectral subtraction, removal of spindle harmonic components, removal of artificial peaks due to aliasing effects, and removal of artificial peaks due to encoder error effects. After filtering, indicator criteria are evaluated to detect chatter, including a magnitude of the filtered torque signal for servo data and a magnitude ratio of the filtered torque signal to the air-cut reference signal for spindle data. Corrective action is taken when chatter is detected.

The invention discloses a cough sound recognition method based on a PSO-GBDT-LR model, and belongs to the technical field of signal recognition. The method includes acquiring audio signals; de-noising the audio signal by using a Berouti spectral subtraction method to obtain a de-noised audio signal; the audio event detection VAD is used for segmenting the part, where the sound appears, of the audio; 7-dimensional time domain features are extracted from each segmented audio sample; performing short-time Fourier transform (STFT) on each segmented audio sample, and extracting two-dimensional frequency domain features from a frequency spectrum; combining the extracted 7-dimensional time domain features and the extracted 2-dimensional frequency domain features to form a 9-dimensional feature vector combination; and marking the feature vector of the cough audio sample as a class 1, and marking the feature vector of the non-cough audio sample as a class 0. According to the method, the problems of excessive noise features and abnormal features of the sound in the cough sound recognition process are solved, the features of the cough and non-cough sound can be accurately distinguished, and the generalization ability is high.

The invention relates to the technical field of agricultural travel services, and provides a dialect intelligent customer service and culture knowledge base system, the system comprises a perception layer, an analysis layer, a knowledge layer and an interaction layer four-dimensional architecture, the perception layer collects and preprocesses dialect voice, noise reduction is performed through spectral subtraction, and Mel frequency cepstrum coefficient features are extracted; the analysis layer identifies dialects based on a fine tuning Wav2Vec2.0 model, converts the dialects into mandarin through a Transform architecture, and completes intention identification and slot filling by using a BERT related model; the knowledge layer constructs a local culture knowledge graph containing non-abandoned, folk and other entities, and supports dynamic updating; and the interaction layer is combined with multiple rounds of dialogue management to generate multiform responses, and can be connected with an external service system. The system also optimizes a feedback module iterative model and knowledge. The system can cover more than ten dialects, realizes dialect interaction, culture interpretation and service conversion closed loop, and assists rural culture revitalizing and rural cultural travel service upgrading.

The invention discloses an optimal linear point searching method for a silicon-based Mach-Zehnder modulator. According to the method, the linearity of a corresponding working point is estimated according to the power of a fundamental component and a third-order intermodulation component of an output signal of the modulator, and the optimal linear point of the modulator can be obtained based on the maximum estimated value of the linearity. According to the method, a maximum entropy spectral subtraction method is adopted to reduce a signal noise floor, and a violent search method is adopted to retrieve an optimal combination of a modulator upper and lower arm phase difference and a reverse bias voltage to obtain an optimal working point. Compared with a conventional linearity test method that the linearity is obtained by testing when the working point of the modulator is set to be an orthogonal point, the optimal linear point searched by the method can obviously improve the linearity of the modulator, and a high-linearity silicon-based integrated microwave photon system can be realized.

The invention discloses a weak fault sound source monitoring and early warning method based on deep learning and an acoustic array, and belongs to the technical field of equipment state monitoring and fault diagnosis. In order to solve the problem of early fault detection caused by easy submergence of weak fault voiceprint signals, incomplete feature extraction and scarcity of fault samples in a strong noise environment, the method comprises the following steps: collecting multichannel signals through an acoustic array, establishing a dynamic noise baseline by adopting a Gaussian mixture model, and performing noise suppression in combination with a spectral subtraction method; extracting a composite feature vector containing time domain, frequency domain and time-frequency domain features; and carrying out fault identification by using the pre-trained and fine-tuned transfer learning model, and finally outputting an early warning according to an identification result. The method is suitable for online monitoring and early warning of early weak faults of industrial equipment.

The invention provides a Transform-based voice correction system and method, and the system comprises a voice preprocessing module which is used for carrying out the noise suppression and feature extraction of an input southern Fujian dialect voice signal, carrying out the noise suppression through a frequency spectrum subtraction method, and extracting 40-dimensional MFCC features; the voice splicing module is used for solving the problem of voice input fragmentation; the dialect speech recognition module adopts an improved Transform encoder structure; and the character correction module adopts a character level Transform language model. According to the invention, more accurate and smooth voice interaction experience is provided for the user in the dialect area in southern Fujian. Meanwhile, the technical framework has good expandability, can adapt to other dialects and language variants, and provides important support for industrial application of the dialect voice technology.

The invention discloses an AI-driven vehicle-mounted sound field real-time modeling and voice separation method, and relates to the technical field of voice signal processing. A main control unit comprising a time sequence synchronizer, a resource scheduler and a health monitor is constructed. 3D sound field modeling is carried out by adopting a lightweight STCN + bidirectional LSTM network, adaptive updating of the model is realized through EWC incremental learning, and a CNN-LSTM noise classification network and targeted suppression algorithms such as ANF / spectral subtraction are developed. The voice separation module adopts an improved Conv-TasNet architecture, 3D spatial constraint and a multi-task loss function are fused, and low delay is realized under INT8 quantization and pipeline processing. The system dynamically optimizes parameters through a real-time regulation and control unit, supports scene self-adaption, finally achieves a separation effect in a mixed noise scene, reduces the delay of the whole system, and effectively improves the definition and stability of vehicle-mounted voice interaction.

The invention relates to the technical field of artificial intelligence, and discloses a BCM module cooperative control method based on voice instruction recognition in a vehicle-mounted high-noise environment, and the method comprises the following steps: collecting original voice instruction data and vehicle state parameter data in the vehicle-mounted high-noise environment; performing voice signal preprocessing in a noise environment based on the original voice instruction data to generate de-noised voice instruction data; according to the method, the engine noise and the wind noise steady-state background noise are filtered out through multi-level noise suppression processing combining adaptive filtering and spectral subtraction, residual noise elimination is carried out for sudden impact noise, and the definition of voice signals is improved. Meanwhile, through a voice activity detection algorithm, a detection threshold is dynamically adjusted according to vehicle state parameters, an effective voice segment and a noise segment can be separated, the accuracy of voice feature extraction in a complex time-varying noise environment is ensured, and thus the robustness and reliability of voice instruction recognition are improved.

The application discloses a consumer complaint response method and system based on AI assistance and language understanding, which splits the complaint response process into two core sub-problems of multi-modal consumer complaint data processing and feature fusion and demand attribution and response generation. In multi-modal consumer complaint data processing and feature fusion, first, regular expressions are used to denoise text, spectral subtraction is used to denoise voice, and Gaussian filtering and adaptive histogram equalization are used to denoise images; then, modal features are extracted, text is used as the core of cross-modal fusion, and entity and relationship are extracted to construct a multi-modal semantic knowledge graph. In demand attribution and response generation, Graph Transformer is used in combination with the graph and domain prior knowledge to output primary and secondary demands; a static complaint graph is constructed, an attribution path is mined through BFS and is verified through multi-modal verification; and an "emotion-demand-attribution-prevention" structure is used to optimize text and adjust the format by using LLM, and an individualized complaint response is output.

This invention discloses a method for accurate audio warning identification based on hybrid features. This method analyzes collected audio speech to determine the issuance time of the warning signal, thereby accurately evaluating the timeliness of the audio warning. First, a double noise reduction method using logmmse-spectral subtraction is employed to filter out noise information in the recorded speech. Next, endpoint detection based on short-time energy is used to mark the effective speech segments in the test speech. Then, MFCC features and waveform polynomial features are extracted from each frame of the effective speech segment. Subsequently, the two features are used as inputs to two channels of a convolutional neural network, and the outputs of the two channels are summed to obtain the hybrid features. Finally, the hybrid features are used as input to a softmax function, and the speech segment containing the target speech (audio warning signal) is determined by the maximum probability value. The starting position of this speech segment is the issuance time of the warning signal.

The application relates to the technical field of artificial intelligence, and discloses a BCM module cooperative control method based on voice instruction recognition in a vehicle high-noise environment, which comprises the following steps: collecting original voice instruction data and vehicle state parameter data in the vehicle high-noise environment; performing voice signal preprocessing in the noise environment based on the original voice instruction data to generate denoised voice instruction data; through multi-level noise suppression processing combining adaptive filtering and spectral subtraction, engine noise, wind noise and steady-state background noise are filtered out, and residual noise is eliminated for sudden impact noise, so that the intelligibility of the voice signal is improved. Meanwhile, through a voice activity detection algorithm, and by dynamically adjusting the detection threshold according to the vehicle state parameters, effective voice segments and noise segments can be separated, the accuracy of voice feature extraction in a complex time-varying noise environment is ensured, and the robustness and reliability of the voice instruction recognition are improved.

The invention relates to the technical field of voice enhancement, in particular to an audio noise reduction method and system for a Bluetooth headset, and the method comprises the steps: analyzing the feature condition of noise influence in a mobile scene, including the feature change condition of a signal obtained by a microphone under the condition of pedestrian voice noise interference; according to the method, the spectral subtraction factor of the spectral subtraction method is adjusted in a targeted manner by integrating the characteristic difference conditions of the audios obtained by the reference microphone and the main microphone when the audios are moved to different scenes, so that the filtering error occurring during the audio noise reduction of the Bluetooth headset in the moving scene is avoided, and the audio noise reduction level in the call process of the Bluetooth headset is further improved.

The invention discloses a distributed reconnaissance tool, relates to the technical field of voice interception, and aims to solve the problems of limited multi-node access capability, poor anti-interference performance and insufficient multi-target interception continuity of an existing reconnaissance system. The tool comprises a micro audio node network and a base station unit, a recording module, a wireless transmission module and an encryption module are arranged in the miniature audio node, local recording and remote transmission are supported, and the miniature audio node has the characteristics of water resistance and low power consumption; the base station unit adopts an orthogonal graph division multiple access technology, at most 32 micro audio nodes can be accessed, four paths of voice can be monitored in real time, the audio quality is optimized through a spectral subtraction voice enhancement algorithm, and WIFI / Ethernet connection and time synchronization and combination of multi-node recording files are also supported. According to the invention, approaching interception and multi-target centralized deployment and control of the moving target are realized, the stability, expansibility and practicability of the interception system are improved, and the method is suitable for scenes such as public security technical investigation and the like.

The invention provides a real-time speech enhancement method and system based on dual-stage spectral subtraction and dual-mask fusion, and relates to the technical field of speech signal processing, and the method comprises the steps: respectively carrying out the framing, windowing and short-time Fourier transform of a left channel mixed signal and a right channel noise reference signal, obtaining a complex frequency spectrum and an amplitude spectrum of the left channel signal and an amplitude spectrum of the right channel noise reference signal; and performing noise estimation by adopting a first noise multiplication factor based on the amplitude spectrum of the right channel noise reference signal to obtain a noise estimation spectrum, and performing constraint spectrum subtraction on the amplitude spectrum of the left channel signal to obtain voice amplitude estimation of a first stage. According to the method, effective suppression of TTS noise and real-time speech enhancement are realized through framing windowing and frequency domain conversion in combination with over-estimation spectrum subtraction and double-mask fusion through two-stage gain application and time domain reconstruction.

The invention relates to the field of float glass production, in particular to a method and a device for identifying glass fragmentation sound in an annealing kiln. According to the method, a plurality of acoustic sensors which are arranged on a longitudinal steel beam on the non-transmission side of the annealing kiln and can tolerate the high temperature of 100 DEG C or above are used for collecting field environment sound signals; sequentially executing time domain preprocessing, improved spectral subtraction noise suppression, wavelet packet analysis and scale energy feature extraction to obtain a normalized feature vector; and finally, carrying out model training and real-time identification based on a hidden Markov model, and outputting whether the sound is glass fragmentation sound or not and a specific fragmentation type. The invention further provides a device for implementing the method, manual guarding can be replaced, accurate recognition and type judgment of glass fragmentation of the annealing kiln are achieved, the labor cost is reduced, manual monitoring defects are avoided, the fragmentation interval can be rapidly positioned, the device is adaptive to the high-temperature and high-noise scene of the annealing kiln, and the yield loss and the equipment production halt risk are reduced.

The electronic processing device for an acoustic apparatus including a first air conduction microphone and a second bone conduction microphone, configured for being connected to the first and second microphones, for receiving as inputs the first and respectively second analog signals from the first, and respectively second microphones and for delivering as output a corrected signal.The processing device comprises:a hybridization module configured for calculating a hybrid signal from the first and second analog signals;an estimation module configured for estimating noise in the hybrid signal;a noise reduction module configured for calculating the corrected signal by applying a generalized spectral subtraction algorithm to the hybrid signal and according to the estimated noise.

This invention relates to a method for highlighting spike discharge data of intracranial electroencephalogram (EEG) signals. The method involves collecting background noise data from the patient's brain without neuronal discharges, preprocessing the background noise data, automatically selecting the optimal order of an autoregressive (AR) model using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), estimating the AR model coefficients using the Yul-Walker equation, and constructing a background noise model. The intracranial EEG signals to be processed are then subjected to high-pass filtering. A short-time Fourier transform (STFT) and window-based frame-by-frame processing strategy are used to subtract the spectrum of the signal from the noise. By adjusting the parameters of the spectral subtraction, noise removal and preservation of neuronal signal features are achieved.

The invention relates to the technical field of hearing aids, in particular to a digital hearing aid howling suppression method and system, and the method comprises the steps: determining an energy distribution feature value; based on the distribution of all high-energy values and derivative values of each to-be-processed signal and a preset number of to-be-processed signals before the to-be-processed signal in the frequency domain and the rate at which each peak value in each to-be-processed signal drops to an adjacent next valley value, determining the noise-containing possibility degree of each to-be-processed signal, and determining the noise-containing possibility degree of each to-be-processed signal in combination with the energy distribution characteristic value. Determining an over-subtraction index; and based on the over-subtraction index, suppressing the howling noise in the voice signal by adopting a spectral subtraction method. According to the method, the frequency domain energy distribution and the time domain oscillation characteristics are dynamically analyzed, the over-subtraction index of the spectral subtraction method is adaptively adjusted, the problem that a traditional howling suppression method is poor in adaptability is solved, and the howling suppression effect of the digital hearing aid in a complex acoustic environment is improved.

The invention relates to an intelligent home switch control system and method based on voice recognition. The intelligent home switch control system comprises a home intranet, a home network block and an external network. According to the smart home switch control system and method based on voice recognition, a home intranet is established by adopting a ZigBee technology, data communication between the home intranet and an OneNET cloud platform extranet is realized by utilizing a WiFi technology, and a user can check real-time environment data and remotely control household appliances through a computer and a mobile phone. Meanwhile, under the condition that a mobile phone cannot be used, voice control over household appliances is achieved through the voice recognition technology, the system is made to be more intelligent, meanwhile, the Berroute spectral subtraction line and the Wiener filter are combined, noise is reduced more efficiently, the sound quality is improved, the voice recognition rate is increased, the federal learning safety module is additionally arranged, and the safety of the system is improved. A federal learning framework is adopted, user voiceprint features are locally encrypted and stored, and original data leakage and replay attack stealing are avoided.

The application provides a speech enhancement method based on deep learning assisted spectral subtraction, relates to the technical field of environmental noise suppression, and aims to improve the explainability of a network in deep learning and achieve excellent noise reduction effect, comprising initial feature extraction, obtaining a log power spectrum feature and a phase feature of a noisy speech signal; obtaining an enhanced amplitude spectrum according to the log power spectrum feature of the noisy speech signal and an estimation and noise reduction network; and obtaining an enhanced time domain speech signal through inverse short-time Fourier transform according to the enhanced amplitude spectrum and the initial phase. The application has the advantages of further improving the speech enhancement quality and the explainability of the network in deep learning.

The invention relates to a non-contact fault detection method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring a motor audio signal for a to-be-tested motor; performing noise suppression processing on the motor audio signal, and extracting to obtain motor audio feature information; the noise suppression processing comprises the steps of dynamically adjusting a noise threshold by utilizing self-adaptive frequency spectrum subtraction, and highlighting the characteristic frequency of the motor through harmonic enhancement processing; and based on the motor audio feature information, utilizing a pre-trained motor fault detection model to obtain a fault classification result for the to-be-detected motor. The reliability of motor fault detection can be effectively improved.