56 results about "Abnormal voice" patented technology

The invention belongs to the technical field of power grid main equipment diagnosis, and provides a power grid main equipment fault diagnosis method and system based on voiceprint cloud edge collaboration, which adopts a new cloud edge collaboration method and divides the analysis of voiceprint monitoring data into two stages. According to the method, the exception detection model is deployed on the edge side, exception analysis of mass voiceprint data is achieved, concurrent requests of a large number of concurrent signals at the cloud side are avoided, and the timeliness of voiceprint monitoring and analysis is improved. In the cloud, a multi-modal defect identification method fusing voiceprint, visible light, infrared and other multi-source data is deployed, secondary classification of sound abnormal signals is realized, and the reliability of equipment operation, maintenance and management is improved.

The invention relates to the technical field of sound anomaly detection, in particular to a sound anomaly detection method and device based on a Transform model, equipment and a medium, and the method comprises the steps: collecting a sound signal during the operation of the equipment through a data collection interface, and obtaining an original sound signal; resampling is carried out on the collected sound signals, and normalization processing is carried out on the resampled data; mel-frequency cepstrum coefficient features are extracted from the sound signals after normalization processing, and the sound signals after normalization processing are input into a pre-training module to output high-dimensional features including time sequence and semantic information; splicing the Mel-frequency cepstrum coefficient features with the high-dimensional features to form a comprehensive feature vector; inputting the comprehensive feature vector into a support vector machine model, and performing abnormal sound recognition through a trained classification hyperplane; and detected abnormal information is fed back to the user in real time. Multi-feature fusion enables the model to identify abnormal sound more accurately.

The invention discloses a transformer fault detection method and device based on an automatic encoder and multi-scale feature fusion, and belongs to the technical field of power transformer fault detection.The method comprises the steps that sound signals during operation of a transformer are collected, and data frame division and normalization preprocessing are conducted on the collected sound signals; reconstructing the data of the collected sound signals by using an automatic encoder trained by normal sound data, and extracting the data coded by the automatic encoder as the implicit features of the sound signals; according to the error before and after reconstruction, sound abnormity judgment is carried out, and fault type detection is further carried out on abnormal data by using a classifier; and the classifier receives the collected sound data and the characteristic signal extracted by the automatic encoder as input, and classifies fault types. According to the method, the classifier is arranged, fault features are mined through the multi-scale features of the transformer sound signals, and the accuracy of transformer fault classification is improved.

The invention provides an equipment fault detection method, a robot and electronic equipment, and the method comprises the steps: obtaining the sound data of target equipment, and carrying out the abnormality prediction of the sound data through a sound abnormality detection model, and obtaining a prediction result; obtaining a visible light image and an infrared image of the target equipment, and determining the real-time temperature of the target equipment according to the visible light image and the infrared image; and performing fault detection on the target equipment according to the prediction result and the real-time temperature of the target equipment. Mutual verification is carried out through combination of sound abnormity of the target equipment and real-time temperature detection determined by an image, fault detection is carried out on the target equipment according to a prediction result of sound data and the real-time temperature of the target equipment, and the multi-modal data of the real-time temperature determined by the sound data and the image can evaluate the health condition of the equipment more comprehensively. The probability of false alarm possibly caused by environmental noise by depending on sound alone is reduced, and the accuracy of fault detection on the equipment is greatly improved.

The invention provides an unsupervised sound anomaly recognition method and device based on a mask strategy, and the method comprises the steps: obtaining normal audio data of machine equipment, and carrying out the preprocessing of the normal audio data, so as to obtain a preprocessed logarithmic Mel spectrogram; an anomaly recognition model of the deep convolutional generative adversarial network is constructed, and the anomaly recognition model comprises a generator, a global discriminator and a random local discriminator; the preprocessed logarithm Mel spectrogram is masked, an anomaly recognition model is trained according to the masked logarithm Mel spectrogram and the preprocessed logarithm Mel spectrogram so as to obtain a trained anomaly recognition model, and in the training process, the logarithm Mel spectrogram and the preprocessed logarithm Mel spectrogram are matched with each other; optimizing the anomaly recognition model by adopting mask loss, reconstruction loss, local feature constraint loss and global feature constraint loss; adopting a trained anomaly recognition model to obtain an anomaly score so as to determine whether the corresponding to-be-recognized sample is abnormal or not; therefore, the sound anomaly detection effect is improved.

The invention provides a power plant equipment sound abnormity identification and health prediction method based on granular computing and an LSTM network, and the method comprises the steps: employing an array pickup and a high-dynamic-range microphone array in a complex and high-noise background environment of a power plant, and combining an anti-interference filtering and beam forming algorithm, thereby achieving the sound abnormity identification and health prediction of the power plant equipment. According to the method, directional, multi-channel and non-contact sound acquisition is carried out on key parts of equipment, acquired sound signals are preprocessed, converted into time domain, frequency domain and time-frequency domain representations and coded into multi-dimensional numerical vectors, and compared with a rule-based expert system, the method has higher generalization ability and real-time response ability; a large language model is introduced, so that the output is closer to an engineering language and is suitable for operation and maintenance personnel to understand and execute; a self-defined knowledge base or safety semantic filtering is supported, and closed-loop deployment in an industrial field is facilitated; the system can be in butt joint with an intelligent inspection system, and full-link linkage of voice recognition, health assessment and suggestion generation is achieved.

The invention discloses a calibration data set acquisition method, a spoken language test scoring method and a related device. The method comprises the following steps: acquiring at least one abnormal voice data; wherein the abnormal voice data is first answering voice data with an abnormal sound event, and the first answering voice data is obtained by performing voice answering on the oral test by an answerer; obtaining a target calibration data set by using the at least one abnormal voice data and the initial calibration data set; wherein the target calibration data set is used for training the oral test scoring system. In this way, the coverage range of the calibration data set can be wider.

The invention relates to the technical field of artificial intelligence, and discloses a sound anomaly detection method and device based on self-supervised learning, equipment and a medium, and the method comprises the steps: obtaining a first sound sample set of mechanical equipment in a normal operation state and a sound sample of to-be-tested mechanical equipment; performing data enhancement processing on the first sound sample set to obtain a second sound sample set; inputting the second sound sample set into the initial encoder neural network model to extract a feature vector; constructing a loss function and updating the encoder neural network model; inputting a sound sample of to-be-tested mechanical equipment into the trained encoder neural network model to extract a feature vector; and calculating the mahalanobis distance between the feature vector of the sound sample of the mechanical equipment to be tested and the feature vector of the second sound sample set, and judging whether the operation state of the mechanical equipment is normal or not according to a calculation result. According to the method, the fault detection accuracy of the industrial mechanical equipment is effectively improved.

The invention provides an abnormal inflexion voice control blocking method and device and an intelligent central control system, and the method comprises the steps: detecting a voice instruction sent by a current user in real time, and obtaining a voice signal; detecting an inflexion feature of the voice signal to judge whether the current user has an inflexion operation or not; under the condition that the current user has the inflexion operation, a multi-verification mechanism is executed to obtain a verification result, the verification result is that the current user is a pre-stored permission user or the current user is a non-permission user, and the multi-verification mechanism is a measure for carrying out user identity verification in multiple modes; and when the verification result is that the current user is the non-authorized user and the voice instruction is executed, adjusting the equipment state to an initial state. The problem that in the prior art, inflexion operation is recognized, but inflexion operation is not effectively stopped, so that the system safety is reduced is solved.

The invention discloses an abnormal sound detection method for fault diagnosis of a belt conveyor. The abnormal sound detection method comprises the following steps: constructing a training set; the training set comprises normal sound and abnormal sound; and training a preset belt conveyor fault diagnosis sound anomaly detection network by using the training set. The preset sound anomaly detection method for fault diagnosis of the belt conveyor mainly comprises a logarithmic Mel spectrogram module, a spectrum-time feature fusion module and a flow-based self-supervision estimation module. Effective features of the operation sound of the belt conveyor in the noisy environment are extracted, fault diagnosis is carried out, and whether a fault occurs or not and the fault type are judged. The logarithmic Mel spectrogram module and the spectral time feature fusion module are fused on the stream-based self-supervised density estimation module, so that multi-scale feature fusion is realized, extracted features are enriched, and an excellent abnormal sound detection effect of belt conveyor fault diagnosis is obtained.

The invention relates to the technical field of sound detection, in particular to a sound anomaly detection smart bracelet and a detection method, and a sound anomaly detection method. The smart bracelet for sound anomaly detection is matched with an analysis host and acoustic detection software, and an acoustic detection algorithm based on spectral analysis and machine learning is used for detecting the sound anomaly. And automatic detection and alarm of abnormal sound are realized. According to the invention, on-line automatic detection of wiring harness plugging is realized, the outdated condition of manual detection of an existing production line is replaced, the detection efficiency and accuracy are higher than those of manual detection, the SOP automation level is improved, and information archiving and backtracking are realized.

The present invention discloses a Whisper audio transcription method, system, and storage medium based on hot word excitation. The present invention proposes a hot word framework, adds a bias encoder / decoder module to the Whisper base framework, and modifies a part of the calculation logic. Among them, the bias encoder is used to extract hot word text features, and the bias decoder combines the hot word feature excitation to transcribe the text, so that the keywords that may have been transcribed incorrectly converge in the direction of the hot word feature. Through hot word excitation, Whisper corrects the erroneous text and obtains the correct text content. Similarly, when an abnormal voice transcription occurs on site, the transcription error correction can be quickly achieved by retaining / inputting hot words, thereby optimizing the fluency of human-computer interaction.

The application discloses an abnormal voice data detection method and device, computer equipment and a storage medium, and applies to the technical field of voice detection. The method comprises the following steps: obtaining historical voice data, obtaining a first corpus set according to the historical voice data, obtaining a second corpus set, and obtaining a representative word set according to the second corpus set and the first corpus set. When the current voice segment of the i-th service scene is detected, whether at least one abnormal word is contained in the current voice segment is determined according to the representative word set corresponding to the i-th service scene before the current voice segment is sent. When at least one abnormal word is contained in the current voice segment, a preset abnormal processing rule is obtained, and the current voice segment is processed based on the preset abnormal processing rule, so that the detection efficiency of abnormal voice data is improved.

The invention provides a self-supervised sound anomaly recognition method and device based on multivariate feature enhancement, and the method comprises the steps: carrying out the preprocessing of original data, carrying out the coding of the attribute information of machine equipment through employing a Mixup technology, generating a classification tag, constructing an auxiliary classification task, and carrying out the recognition of the classification tag; the method comprises the following steps: respectively extracting a logarithmic Mel spectrogram, a speech spectrogram and a spectrogram of audio data as the input of a multivariate feature extraction network, aggregating the output of the multivariate feature extraction network to serve as the feature of a single audio sample, and enhancing the discrimination and robustness of the feature by using a multivariate feature enhancement method based on cross fusion, so as to improve the robustness of the audio sample. Finally, the features of the feature space are clustered, and a clustering center is generated; in the test stage, to-be-tested data is input into the feature extraction network to extract features, and an exception score is calculated by comparing cosine distances between the features, so that exception is judged; therefore, the sound anomaly detection effect is improved.

To provide an output sound abnormality detecting device capable of detecting an abnormality in an output system including a content of an output sound.SOLUTION: An on-vehicle device 100 comprises: a warning sound generation unit 24 that generates an L channel warning sound that is an output target and an R channel sound having the same content as the warning sound; an amplifier 30 that amplifies these sounds and outputs an amplified sound; a comparator 40 that compares signal levels of the two sounds output from the amplifier 30; and a warning sound abnormality determining unit 28 that determines whether or not there is an abnormality in the warning sound and the amplifier 30 based on a comparison result by the comparator 40.SELECTED DRAWING: Figure 1

The invention discloses a voice interaction method and device, equipment and a storage medium, and relates to the technical field of voice processing, and the method comprises the steps: determining an optimal voice collection mode of a voice signal under the condition that the voice signal of a user is detected; acquiring a real-time voice signal of the user through the optimal voice acquisition mode; detecting whether abnormal voice signals exist in the real-time voice signals or not, wherein the abnormal voice signals are voice signals with pronunciation problems; if yes, repairing the real-time voice signal; and generating and executing a voice control instruction according to the restored voice signal. According to the invention, when the real-time voice signal of the user has the pronunciation problem, the real-time voice signal can be repaired, and the voice control instruction is generated and executed according to the repaired voice signal, so that the problem that the existing far-field voice recognition system cannot accurately recognize the voice instruction with the voice problem is solved, and the voice recognition efficiency is improved. And the voice interaction is limited.

This invention relates to a drive control method and system for a micro motor, belonging to the field of intelligent control. The method includes: acquiring real-time sound signals and the motor model during micro motor operation; obtaining a reference sound signal based on the motor model; comparing the real-time sound signal with the reference sound signal to determine if there is a sound anomaly in the micro motor; if a sound anomaly is determined, obtaining abnormal acoustic signature features and sound acquisition location information based on the real-time sound signal; combining the abnormal acoustic signature features and sound acquisition location information to locate the specific anomaly location; matching the anomaly handling method according to the specific anomaly location and abnormal acoustic signature features, and controlling a preset actuator to perform the corresponding processing operation according to the anomaly handling method, while simultaneously reporting a sound anomaly handling prompt. This application has the effect of improving the timeliness of fault early warning.

The invention relates to the field of artificial intelligence, and discloses a sound anomaly detection method and device based on self-supervised learning, electronic equipment and a storage medium, and the method comprises the steps: carrying out the voice enhancement of a positive sample sound data set, and obtaining an enhanced sample sound data set; training a preset feature extraction model by using the positive sample sound data set and the enhanced sample sound data set to obtain a trained feature extraction model; extracting features of the positive sample sound data set and the to-be-detected sound data by using the trained feature extraction model to obtain a standard sound wave feature data set and to-be-detected sound wave feature data; calculating a mean value and a covariance of the sound wave characteristic data set to obtain a standard sound wave mean value and a standard sound wave covariance; according to the standard sound wave mean value and the standard sound wave covariance, the distance between the to-be-detected sound wave feature data and the sound wave feature data set is calculated, and when the distance is larger than a preset threshold value, it is judged that the to-be-detected sound data is abnormal data. According to the invention, the accuracy of sound anomaly detection can be improved.

The present application provides a kind of traditional Chinese medicine preparation for treating Tourette syndrome of children with laryngeal abnormal voice as main symptom, the traditional Chinese medicine preparation includes the following weight parts of raw medicinal materials: 8-15 parts of radix isatidis, 3-6 parts of radix sophorae tonkinensis, 10-15 parts of radix puerariae, 6-10 parts of jinyu longan, 6-10 parts of momordica grosvenori, 6-8 parts of platycodon grandiflorum, 3-6 parts of licorice, 5-8 parts of periostracum cicada, 10 parts of radix angelicae, and addition and subtraction according to different indications, the traditional Chinese medicine preparation of the present application can clear lung and stomach heat, benefit throat, block the way of external wind driving internal wind, can rapidly and effectively treat pediatric vocal tic, improve laryngeal voice symptoms, also can treat exogenous heat syndrome sore throat, and the traditional Chinese medicine preparation of the present application is safe, no obvious toxic side effects, suitable for children to take when laryngeal abnormal voice.

To provide a device that processes acoustic data detected in a facility to identify an apparatus in which an abnormality occurs.SOLUTION: The device includes: an acoustic data acquisition section that acquires acoustic data detected by a robot capable of moving in a facility and collecting sound; a first sound abnormality detection section that detects an abnormal sound in a first direction centering around the robot using the acoustic data; a second sound abnormality detection section that detects an abnormal sound in a second direction centering around the robot using the acoustic data; and an abnormal apparatus identification section that identifies an apparatus in which an abnormality occurs using a position of the abnormal sound in the first direction and a position of the abnormal sound in the second direction.SELECTED DRAWING: Figure 10

A method for detecting anomalies has the following steps:Obtaining a long-term recording having a plurality of first audio segments associated to respective first time windows; analyzing the plurality of the first audio segments to obtain, for each of the plurality of the first audio segments, a first characteristic vector describing the respective first audio segment; obtaining a further recording having one or more second audio segments associated to respective second time windows; analyzing the one or more second audio segments to obtain one or more characteristic vectors describing the one or more second audio segments ABCD; matching the one or more second characteristic vectors with the plurality of the first characteristic vectors to recognize at least one anomaly, like a temporal, sound or spatial anomaly.

The application relates to the field of artificial intelligence and discloses a sound anomaly detection method and device based on self-supervised learning, an electronic device and a storage medium. The method comprises the following steps: performing speech enhancement on a positive sample sound data set to obtain an enhanced sample sound data set; training a preset feature extraction model by using the positive sample sound data set and the enhanced sample sound data set to obtain a trained feature extraction model; extracting features of the positive sample sound data set and to-be-detected sound data by using the trained feature extraction model to obtain a standard sound wave feature data set and to-be-detected sound wave feature data; calculating the mean and covariance of the sound wave feature data set to obtain a standard sound wave mean and a standard sound wave covariance; and calculating the distance between the to-be-detected sound wave feature data and the sound wave feature data set according to the standard sound wave mean and the standard sound wave covariance, and determining that the to-be-detected sound data is abnormal data when the distance is greater than a preset threshold. The application can improve the accuracy of sound anomaly detection.

The invention discloses a speech recognition method, device and system. Receiving a voice slice sent by the CCU through the websocket connection, and caching the voice slice; in response to the arrival time of a preset duration period, splicing all the cached voice slices into a complete long voice according to a caching sequence; sending the long voice to an online voice recognition server, enabling the online voice recognition server to perform voice recognition on the long voice to obtain a transcription text, and returning the transcription text; receiving a transliteration text returned by the online voice recognition server, sending the transliteration text to a large language model server, enabling the large language model server to judge the transliteration text through a preset large language model, obtaining a corresponding judgment result, and returning the judgment result; and in response to determining that the long voice is abnormal voice according to the judgment result, stopping voice slice splicing, and returning the judgment result to the CCU, so that the CCU performs processing according to a preset abnormality processing strategy. According to the method, voice recognition can be accurately realized in real time.

The invention discloses an automatic audio device self-checking method based on embedded telephone equipment, and relates to the field of audio signal processing and hardware diagnosis. The method comprises the steps that firstly, a microphone collects first 3S playback signals of a test audio, and frequency similarity detection is carried out by extracting spectrum features; then, continuously collecting the 4s-5s playback signal of the audio, and carrying out single frequency point detection; meanwhile, aiming at the front 3S playback signals, calculating a root-mean-square RMS value according to signal energy, converting the RMS value into loudness, and carrying out loudness anomaly detection; thirdly, continuously collecting the 6s-7s playback signal of the audio for sampling analysis, calculating the root-mean-square of the playback signal, converting the root-mean-square into loudness L2, and when the loudness value L2 is lower than a set threshold value, indicating that the MIC device is abnormal; and finally, through frequency similarity detection, single-frequency detection, loudness anomaly detection and device anomaly detection, results of four rounds of detection are reported in a binary form, and equipment sound anomaly reasons are analyzed. According to the invention, specific abnormal links in the audio link can be effectively positioned.

The invention relates to the technical field of audio detection, in particular to a sound abnormity monitoring processing method and system in the operation process of industrial equipment, and the method comprises the steps: carrying out the data packet segmentation of equipment audio data generated in the operation process, and configuring and outputting a state conversion diagram associated with the equipment audio data; based on the state transition diagram, performing data packet signal recombination between energy sound channel audio data packets in the equipment audio data, and outputting a plurality of reference audio data associated with the equipment audio data; for each reference audio data, based on the fundamental tone frequency data, calculating the anomaly detection confidence of each energy sound channel audio data packet; and performing anomaly detection on the industrial equipment operation audio library based on the anomaly detection confidence of each energy sound channel audio data packet, and outputting an anomaly detection result associated with the equipment audio data. According to the invention, richness and diversity of data packet expression of the equipment audio data caused by different energy positions in the equipment audio data are fully considered, and the accuracy of anomaly detection is improved.

The invention discloses a voice information processing method and device, a storage medium and electronic equipment, and relates to the field of artificial intelligence. The method comprises the following steps: converting target voice information of a user into target text information; inputting the target text information into a target model, determining a feature vector of the target text information according to priori knowledge learned by the target model in a model training stage, and generating a probability vector for the target voice information based on the feature vector; a target probability value and a target dimension corresponding to the target probability value in the probability vector are determined, and the target probability value is larger than a first preset threshold value; and determining whether the target voice information is abnormal voice according to the target probability value, the target dimension and a second preset threshold. According to the method and the device, the technical problem that the false alarm rate is high when illegal committed sales records are recognized through keyword matching in the sales process of financial products is solved.

The invention relates to the technical field of sound analysis, in particular to a throat lesion monitoring system based on sound recognition, which comprises a sound quality optimization module, a sound feature analysis module, an abnormality evaluation module and a treatment effect evaluation module. According to the invention, by utilizing the historical sound data of the user and comparing and evaluating the deviation of the current sound features, the monitoring accuracy is improved, the detection of early lesions is more reliable, the deviation detection threshold is adjusted according to the age and gender differences of individuals, the sound anomaly detection is more accurate, the sound anomaly of the user can be found in time, and the user experience is improved. The system has the advantages that the system is simple in structure and convenient to use, false alarm and missing alarm are reduced, real-time and visual treatment feedback is provided by collecting sound data before and after treatment, analyzing the sound change degree, combining the treatment time and evaluating the treatment effect, data support is provided for doctors, measure adjustment in the treatment process is optimized, and the disease monitoring and management fineness and effect are improved.

The present application discloses a speech evaluation method, which includes: obtaining a speech to be evaluated and a reference text, extracting acoustic features from the speech, inputting the acoustic features into K sub-acoustic models of an acoustic model, and obtaining a comprehensive state posterior probability; aligning the phonemes of the speech with the reference text according to the comprehensive state posterior probability and a segmentation network constructed by the reference text, and obtaining a scoring feature based on the alignment result; and inputting the scoring feature into a scoring model to regress the score of the speech. Since the K sub-acoustic models are trained by their own in-set data and out-of-set data, and the out-of-set data of each sub-acoustic model is different, the acoustic model has a distinguishing feature for the out-of-set data, thereby improving the scoring accuracy of abnormal speech, solving the problem of unstable scoring of abnormal speech, and having high reliability and availability.

The invention belongs to the technical field of artificial intelligence and intelligent security and protection, and particularly relates to a campus hidden area abnormal voice sensing and classification recognition system. Comprising a distributed acoustic sensing array, an environment context sensing module, an adaptive signal conditioning module, a sound source credibility evaluation module, a context enhanced abnormal voice classification module and a privacy protection type data processing module. Through a multi-mode environment perception and dynamic feedback coupling mechanism, sound source space verification, adaptive gain control and multi-level classification decision are realized, and through an adaptive pickup mechanism driven by an environment context, excessive acquisition of environment noise in an unmanned state is avoided; therefore, false triggering caused by non-human-body sound sources can be eliminated by utilizing consistency verification of sound source spatial orientation and human-body existence information.

The invention discloses a bipartite graph clustering-based live pig sound anomaly detection method and device, and the method comprises the steps: carrying out the preprocessing of collected audio data, and obtaining a pig sound standardized audio sample; extracting acoustic characteristic parameters of the audio samples, and constructing a sample characteristic matrix; dividing the sample feature space into a plurality of balanced sub-clusters by adopting a balanced hierarchical K-means algorithm, and selecting a centroid as an anchor point; the sample-anchor point similarity is calculated, and a bipartite graph is constructed; inputting the bipartite graph into a preset clustering model BGCB, optimizing a sample and an anchor label matrix by adopting an alternating coordinate descent algorithm, generating a clustering result and calculating a clustering center; and identifying abnormal sounds of pigs such as cough, squeal and moan according to the clustering center difference and the spectrum energy distribution characteristics, and outputting an alarm. The method can achieve the automatic detection of the abnormal sound of the live pigs without manual marking, has the advantages of balanced clustering result, low calculation complexity and high robustness, and is suitable for the health monitoring and disease early warning of intelligent farms.