A noise detection analysis method and system

Abstract

The application provides a noise detection analysis method and system, which is suitable for the technical field of data processing, and comprises the following steps: sampling, feature extraction and standardization processing of a plurality of to-be-detected sound signals to obtain a plurality of to-be-detected sound feature information; classifying the plurality of to-be-detected sound feature information to generate a plurality of sound detection category information; and detecting and analyzing the plurality of sound detection category information to generate a plurality of noise detection analysis information. The application improves the spatial resolution of noise source positioning, enhances the classification accuracy of the plurality of sound detection category information and the adaptability to complex noise patterns, thereby improving the detection analysis accuracy and effectiveness of noise.

Description

Technical Field

[0001] This application belongs to the field of data processing technology, and in particular relates to noise detection and analysis methods and systems. Background Technology

[0002] In the field of audio signal processing and noise detection and analysis, with the continuous expansion of application scenarios such as electronic devices and industrial machinery, the demand for audio signal processing accuracy and noise detection efficiency continues to rise.

[0003] In existing technologies, the acquired sound signal is usually sampled first, discretized at a fixed frequency, and then frequency domain features are obtained by relying on traditional Fourier transform, or noise in the sound signal is identified by using simple time-domain statistical feature calculation methods.

[0004] However, in existing technologies, fixed-frequency sampling is difficult to adapt to the complex and ever-changing characteristics of sound signals. It is difficult to accurately capture abrupt signals or sounds containing rich details. Feature extraction cannot fully and deeply mine the effective features in the sound signal, resulting in the omission of key information and difficulty in effectively distinguishing similar noise categories. Summary of the Invention

[0005] In view of this, embodiments of this application provide a noise detection and analysis method and system, which aims to solve the problems of lack of adaptability of sampling methods, insufficient feature extraction, and single detection and analysis dimensions in the prior art.

[0006] The first aspect of this application provides a noise detection and analysis method, including:

[0007] Acquire multiple sound signals to be detected;

[0008] The multiple sound signals to be detected are sampled, feature extracted, and standardized to obtain multiple sound feature information to be detected.

[0009] The multiple sound feature information to be detected is classified to generate multiple sound detection category information;

[0010] The multiple sound detection category information is processed and analyzed to generate multiple noise detection analysis information.

[0011] A second aspect of this application provides a noise detection and analysis system, including:

[0012] The sound signal acquisition module is used to acquire multiple sound signals to be detected;

[0013] The detection sound feature information generation module is used to sample, extract features and standardize the multiple detection sound signals to obtain multiple detection sound feature information;

[0014] The sound detection category information generation module is used to classify the multiple sound feature information to be detected and generate multiple sound detection category information.

[0015] The noise detection and analysis information generation module is used to detect, analyze, and process the multiple sound detection category information to generate multiple noise detection and analysis information.

[0016] A third aspect of this application provides a terminal device, the terminal device including a memory and a processor, the memory storing a computer program executable on the processor, the processor executing the computer program to implement the steps of the noise detection and analysis method described in the first aspect above.

[0017] A fourth aspect of this application provides a computer-readable storage medium, comprising: storing a computer program, which, when executed by a processor, implements the steps of the noise detection and analysis method described in the first aspect above.

[0018] The beneficial effects of this application embodiment compared with the prior art are: by constructing a comprehensive acoustic sensing network, this application improves the accuracy of classifying multiple sound signals and enhances the recognition effect of complex noise patterns, thereby effectively enhancing the accuracy and effectiveness of noise detection, analysis and source tracing. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment 1 of this application;

[0021] Figure 2 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment 2 of this application;

[0022] Figure 3 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment 3 of this application;

[0023] Figure 4 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment 4 of this application;

[0024] Figure 5 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment 5 of this application;

[0025] Figure 6 This is a schematic diagram of the implementation process of the noise detection and analysis method provided in Embodiment Six of this application;

[0026] Figure 7 This is a schematic diagram of the structure of the noise detection and analysis system provided in the embodiments of this application;

[0027] Figure 8 This is a schematic diagram of the terminal device provided in the embodiments of this application. Detailed Implementation

[0028] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0029] To illustrate the technical solution described in this application, specific embodiments are provided below.

[0030] Figure 1 A flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment 1 of this application is shown, and is described in detail below:

[0031] Step S101: Acquire multiple sound signals to be detected.

[0032] In this embodiment, the sound signal to be detected can refer to multiple independent sound data sequences that need to be analyzed for noise detection, collected synchronously or in time-division by acoustic sensing devices in the acoustic environment related to muffler manufacturing and industrial equipment operation environment. These sequences can include raw acoustic information from various sound sources in the muffler manufacturing scenario, such as equipment operating sounds, ambient background sounds, and interference noise. The signal is in the form of a continuous analog sound wave signal, which needs to be converted into calculable discrete data through digital processing. The source scenario of the sound signal to be detected can be clearly defined based on actual needs such as muffler performance testing and equipment noise monitoring, and the sound source range to be covered can be delineated to ensure that the collected sound signal to be detected can comprehensively reflect the acoustic characteristics within the scenario. Multiple microphones or acoustic sensors can be deployed in the muffler manufacturing scenario to form a distributed acquisition network. The number of sensors can be determined according to the complexity of the scenario and analysis requirements. The location distribution needs to cover key sound source points and potential noise propagation paths, such as the air inlet, air outlet, and shell surface of the muffler, to ensure that each sensor independently collects the sound signal at its location, thereby achieving the acquisition of multiple sound signals to be detected.

[0033] Step S102: Sample, extract features and standardize the multiple sound signals to be detected to obtain multiple sound feature information to be detected.

[0034] In this embodiment, a suitable sampling frequency can be determined manually first. This frequency needs to be set according to the frequency range that the sound signal to be detected may contain, so as to ensure that the key information in the sound can be completely preserved. According to the determined sampling frequency, each sound signal to be detected is converted from a continuous analog signal into a discrete digital signal. After sampling, each sound signal to be detected forms a sequence composed of a series of discrete values, thereby completing the sampling processing of the multiple sound signals to be detected. In feature extraction, for each sampled sound signal to be detected, a time window segmentation method can be used to divide the long signal into multiple consecutive short time periods, each of which is a window. A certain overlap is set between windows to avoid information loss. Time-domain features are extracted from the signal within each window, including calculating the average amplitude of the signal within the window, the dispersion of the amplitude, the total signal energy, and the number of times the signal crosses zero per unit time. When extracting frequency-domain features, the time-domain signal within the window is converted to the frequency domain to obtain the distribution of the signal at different frequencies. The frequency points with the most concentrated energy, the weighted average of the frequencies, and the proportion of energy within a specific frequency range to the total energy are calculated. When extracting time-frequency domain features, wavelet transform is used to decompose the signal within the window into combinations of different frequency components and time positions. The energy of each decomposed part is calculated. All features extracted from each window are integrated to form the feature set corresponding to that window. Then, the feature sets of all windows of a sound signal to be detected are summarized to obtain the sound feature information corresponding to that sound signal. After the above processing, multiple sound feature information is formed for all sound signals to be detected. Then, for each feature contained in the multiple sound feature information to be detected, the average value and dispersion of the feature in all sound feature information to be detected are calculated. Based on the calculated average value and dispersion, the feature in each sound feature information to be detected is adjusted so that the distribution of the feature in all sound feature information to be detected is at a uniform scale, eliminating the influence of different features due to different dimensions, thereby obtaining the standardized multiple sound feature information to be detected.

[0035] Step S103: Classify the multiple sound feature information to be detected to generate multiple sound detection category information.

[0036] In this embodiment, multiple sound feature information to be detected may first be grouped. A suitable number of groups is determined, which can be manually set based on the complexity of the sound feature information and actual classification requirements. A corresponding number of feature information is selected from the multiple sound feature information to be detected as initial group centers. The similarity between each sound feature information to be detected and each initial group center is calculated. Each sound feature information to be detected is assigned to the group containing its most similar group center. Then, the average feature of all sound feature information to be detected within each group is recalculated as a new group center. This process of calculating the similarity between feature information and group centers, grouping, and updating group centers is repeated until the change in group centers is less than a set threshold, thus completing the grouping of the multiple sound feature information to be detected and obtaining multiple feature groups. Then, the grouped feature information is processed to capture the correlation between features. Three different feature matrices are generated through transformation, representing the information that the current feature needs to focus on, the feature's own attributes, and the specific content of the feature, respectively. These three matrices are split into multiple sub-parts, and the correlation between the information of interest and the attribute is calculated for each sub-part. The specific content is then weighted and integrated based on the correlation degree to obtain the output result of each sub-part. The output results of all sub-parts are then integrated and transformed to obtain the enhanced feature. This feature highlights the key connections between different sound feature information to be detected. The enhanced feature can then be input into a neural network for classification. The neural network contains multiple hidden layers, which progressively map and abstract the input features through nonlinear transformations to extract higher-level feature representations. In the output layer of the network, the probability of each category is calculated, mapping the feature information to preset sound categories. Each sound feature information to be detected corresponds to the category with the highest probability, generating the sound detection category information corresponding to that sound feature information, thus generating multiple sound detection category information.

[0037] Step S104: Detect and analyze the multiple sound detection category information to generate multiple noise detection and analysis information.

[0038] In this embodiment, the basic acoustic parameters corresponding to each sound detection category information can be extracted first. These parameters may include the intensity, duration, and main frequency range of the sound category. By analyzing the waveform and spectrum of the sound signal, the specific values ​​of these parameters are determined, forming preliminary analysis data. Then, combined with the classification results of the sound detection category information, the occurrence patterns of different sound categories are analyzed. The distribution of the frequency and duration of each sound category in different time periods, as well as the coexistence relationships between different sound categories, are statistically analyzed. For example, which categories frequently occur simultaneously, and which categories exhibit obvious alternating characteristics. Finally, the obtained analysis data is... The basic parameters and occurrence patterns are compared with preset noise evaluation standards. Based on the comparison results, it is determined whether each sound detection category belongs to noise and what level of noise it belongs to, such as slight interference noise, moderate impact noise, or severe pollution noise. At the same time, the potential impact of this type of noise is analyzed, such as potential interference with equipment operation and the degree of impact on the environment. Then, all the analysis content is integrated to generate a comprehensive report for each sound detection category, which includes basic parameters, occurrence patterns, noise level, and impact assessment. These reports serve as multiple noise detection analysis information, thus comprehensively presenting the noise characteristics and related information of each sound detection category.

[0039] The noise detection and analysis method provided in this application improves the accuracy of classifying multiple sound signals and enhances the recognition effect of complex noise patterns by constructing a comprehensive acoustic sensing network, thereby effectively enhancing the accuracy and effectiveness of noise detection, analysis and source tracing.

[0040] Figure 2 The flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment 2 of this application is shown. The difference between this method and Embodiment 1 is that step S102 specifically includes:

[0041] Step S201: Sample the multiple sound signals to be detected to obtain multiple sound signal sequence information.

[0042] In this embodiment, a suitable sampling frequency is determined based on the frequency range that the multiple sound signals to be detected may contain, so as to ensure that the key information in the sound can be completely preserved. Each sound signal to be detected can be converted from a continuous analog signal into a discrete digital signal according to the determined sampling frequency. After sampling, each sound signal to be detected forms a sequence composed of a series of discrete values. These sequences are the sequence information of multiple sound signals to be detected.

[0043] Step S202: Based on the preset temporal segmentation interval of the sound signal to be detected, the multiple sound signal sequence information to be detected is segmented to obtain multiple sound signal sub-sequence information to be detected.

[0044] In this embodiment, the preset temporal segmentation interval of the sound signal to be detected can be a fixed time length set according to the characteristics of the sound signal to be detected and the analysis requirements. At the same time, the overlap ratio between the intervals is set to avoid losing important information during the segmentation process. According to the preset temporal segmentation interval, each sound signal sequence information to be detected is divided into multiple continuous and overlapping short time interval sequences. Each short time interval sequence is a subsequence information of the sound signal to be detected, thereby obtaining multiple subsequence information of the sound signal to be detected.

[0045] Step S203: Extract time-frequency domain features from the multiple sub-sequence information of the sound signals to be detected, and obtain time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected.

[0046] In this embodiment, for each subsequence of the sound signal to be detected, time-domain features are extracted, including calculating the average amplitude of the signal within the subsequence, the dispersion of the amplitude, the total signal energy, and the number of times the signal crosses zero per unit time. These features constitute multiple time-domain feature information of the sound signal to be detected. Simultaneously, frequency-domain transformation is performed on each subsequence of the sound signal to be detected to obtain the signal distribution at different frequencies. The frequency points with the most concentrated energy, the weighted average of the frequencies, and the proportion of energy within a specific frequency range to the total energy are calculated. These features constitute multiple frequency-domain feature information of the sound signal to be detected.

[0047] Step S204: Standardize the time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected to obtain multiple sound feature information to be detected.

[0048] In this embodiment, for each time-domain feature contained in the time-domain feature information of multiple audio signals to be detected, the average value and dispersion of that time-domain feature are calculated across all the time-domain feature information of the audio signals to be detected. Based on the calculation results, the time-domain feature in each audio signal to be detected is adjusted. Simultaneously, the same calculation and adjustment are performed on each frequency-domain feature contained in the frequency-domain feature information of multiple audio signals to be detected. The standardized time-domain feature information and frequency-domain feature information of the multiple audio signals to be detected are then integrated to form multiple audio feature information to be detected.

[0049] The noise detection and analysis method provided in this application accurately captures the detailed features of multiple sound signals to be detected, improves the pertinence and effectiveness of feature extraction, provides a more reliable basis for subsequent classification processing, and thus enhances the accuracy and reliability of noise detection and analysis.

[0050] Figure 3The flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment 3 of this application is shown. The difference between this method and Embodiment 1 is that step S103 specifically includes:

[0051] Step S301: Based on the preset number of sound detection categories, randomly extract the multiple sound feature information to be detected to obtain multiple core feature information of the sound to be detected.

[0052] In this embodiment, the preset number of sound detection categories can be set manually or determined based on actual noise detection needs and the complexity of multiple sound feature information to be detected, thus specifying the number of sound categories to be divided. According to this preset number, a corresponding number of feature information is randomly selected from the multiple sound feature information to be detected. These selected feature information constitute the core feature information of the multiple sounds to be detected and will serve as the initial reference benchmark for subsequent classification.

[0053] Step S302: Calculate the logical distance between the multiple sound feature information to be detected and the multiple sound core feature information to be detected, and obtain the classification distance information of the multiple sound features to be detected.

[0054] In this embodiment, the logical distance can be Euclidean distance. For each sound feature to be detected, the logical distance between it and each core feature to be detected is calculated. This distance is used to measure the similarity between the two in the feature space. By calculating the logical distances between multiple sound feature pieces and multiple core feature pieces one by one, all the obtained logical distance results are the classification distance information of multiple sound features to be detected. This information reflects the degree of correlation between each sound feature and the core feature.

[0055] Step S303: Based on the classification distance information of the multiple sound features to be detected and the core feature information of the multiple sound features to be detected, classify the multiple sound feature information to be detected to obtain multiple sound detection category information.

[0056] In this embodiment, each sound feature to be detected is assigned to the category of the core sound feature with the smallest logical distance. After the initial classification, the average feature of all sound features to be detected within each category is recalculated and used as the new core sound feature. The logical distance between the multiple sound features to be detected and the new core sound feature is calculated again. This process of classification and core feature update is repeated until the change in the core feature is less than a set threshold. The resulting category is then the multiple sound detection category information.

[0057] The noise detection and analysis method provided in this application effectively improves the efficiency and stability of classifying multiple sound feature information to be detected, providing a more accurate basis for subsequent noise detection and analysis, thereby enhancing the reliability of noise detection and analysis.

[0058] Figure 4 The flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment 4 of this application is shown. The difference between this method and Embodiment 3 above is that step S303 specifically includes:

[0059] Step S401: Extract the minimum value of the classification distance information of multiple sound features corresponding to the multiple core feature information of the sound to be detected, and obtain the maximum and minimum value information of the classification distance of multiple sound features to be detected.

[0060] In this embodiment, for each core feature information of the sound to be detected, the distance information with the smallest value is found from the multiple classification distance information of the corresponding sound features to be detected. These minimum distance information are the maximum and minimum values ​​of the classification distance of the multiple sound features to be detected, which represent the distance between each core feature information of the sound to be detected and the most similar sound feature information.

[0061] Step S402: Based on the maximum and minimum distance information of multiple sound features to be detected and the core feature information of the sound to be detected corresponding to the maximum and minimum distance information of multiple sound features to be detected, classify the sound feature information corresponding to the maximum and minimum distance information of multiple sound features to be detected to obtain multiple sound detection feature subclass information.

[0062] In this embodiment, each detection sound feature classification distance maximum value information corresponds to a detection sound feature information and a detection sound core feature information. Detection sound feature information corresponding to the same detection sound core feature information is grouped into one category, forming multiple sound detection feature subclass information. Each sound detection feature subclass information is associated with a specific detection sound core feature information.

[0063] Step S403: Calculate the mean of multiple sound detection feature subclass information to obtain multiple sound detection subclass core feature information.

[0064] In this embodiment, for each sound detection feature subclass, the average feature of all sound feature information to be detected in that subclass is calculated. These average features are the core feature information of multiple sound detection subclasses, and they are the representative features of each sound detection feature subclass.

[0065] Step S404: Determine whether the core feature information of the plurality of sound detection subclasses is the same as the core feature information of the plurality of sounds to be detected; if yes, proceed to step S405; if no, proceed to step S406.

[0066] In this embodiment, the core feature information of multiple sound detection subclasses is compared with the core feature information of multiple sounds to be detected to check whether the two are completely consistent in terms of features. If all corresponding features are the same, they are determined to be the same; if any corresponding feature is different, they are determined to be different.

[0067] Step S405: Based on the multiple sound detection feature subclass information, obtain multiple sound detection category information.

[0068] In this embodiment, when multiple sound detection subclass core feature information is the same as multiple sound core feature information to be detected, it indicates that the classification has been stable. At this time, the multiple sound detection feature subclass information is the final determined classification result. Each sound detection feature subclass information corresponds to a sound detection category information, thereby obtaining multiple sound detection category information.

[0069] Step S406: The core feature information of the multiple sound detection subclasses is used as the core feature information of multiple sounds to be detected, and the process returns to step S302.

[0070] In this embodiment, when the core feature information of multiple sound detection subclasses is different from the core feature information of multiple sounds to be detected, the core feature information of multiple sound detection subclasses replaces the original core feature information of multiple sounds to be detected. Then, the logical distance between the core feature information of multiple sounds to be detected and the new core feature information of multiple sounds to be detected is calculated again, and a new round of classification process begins.

[0071] The noise detection and analysis method provided in this application improves the accuracy and stability of classification, ensures that the obtained multiple sound detection category information is more reliable, provides a more solid foundation for subsequent noise detection and analysis, and thus enhances the overall effect of noise detection and analysis.

[0072] Figure 5 The flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment 5 of this application is shown. The difference between this method and Embodiment 1 is that step S104 specifically includes:

[0073] Step S501: Encode and normalize the multiple sound detection category information to obtain multiple sound detection category vectors.

[0074] In this embodiment, the encoding process refers to converting the classification labels, acoustic parameters, and other features in each sound detection category information into digital representations to form initial feature vectors. Subsequently, these vectors are normalized by calculating the norm or statistical feature distribution of the vectors to map all vectors to a uniform numerical range, eliminating dimensional differences, ensuring comparability between features, and ultimately obtaining multiple standardized sound detection category vectors.

[0075] Step S502: Based on multiple preset noise category detection vectors, perform detection calculations on the multiple sound detection category vectors to obtain multiple noise category detection information.

[0076] In this embodiment, the preset noise category detection vectors are predefined standard feature templates representing different noise categories. The degree of matching between the sound detection category and the preset noise category is evaluated by calculating the similarity or distance metric (such as cosine similarity, Euclidean distance, etc.) between each sound detection category vector and each noise category detection vector. These matching results are then aggregated to form multiple noise category detection information sets, each reflecting the probability that the corresponding sound detection category belongs to a specific noise category.

[0077] Step S503: Based on the multiple noise category detection information and the preset noise category detection weight information, obtain multiple noise category detection weight information.

[0078] In this embodiment, the preset noise category detection weight information is used to reflect the differences in importance between different noise categories. Its construction needs to consider factors such as the degree of noise hazard and the priority requirements of the detection scenario. Specific values ​​are determined through statistical analysis and expert evaluation. Common noise categories, such as mechanical vibration noise, electromagnetic interference noise, and airflow noise, can be divided into multiple hazard levels according to their degree of interference with equipment operation and their pollution level to the environment. For example, mechanical vibration noise may cause equipment wear and tear, so its hazard level is set to "high"; airflow noise mainly affects environmental comfort, so its hazard level is set to "medium"; and electromagnetic interference noise may interfere with electronic device signals, so its hazard level is set to "high". Then, basic weight coefficients are assigned according to the hazard level, such as a weight coefficient of 0.4 for "high", 0.3 for "medium", and 0.2 for "low". If there are special requirements for specific scenarios, such as focusing on airflow noise in muffler manufacturing, the weights of specific categories can be adjusted. For example, the weight of airflow noise can be increased from 0.3 to 0.35, and then normalization is performed to ensure that the sum of the weight coefficients of all noise categories is 1. For instance, if the initial weights of mechanical vibration noise, electromagnetic interference noise, and airflow noise are 0.4, 0.4, and 0.35 respectively, with a sum of 1.15, then by scaling proportionally, the final weights can be obtained by dividing each weight by 1.15: mechanical vibration noise 0.35, electromagnetic interference noise 0.35, and airflow noise 0.3, with a sum of 1. The final preset noise category detection weight information is: mechanical vibration noise 0.35, electromagnetic interference noise 0.35, and airflow noise 0.3, used to highlight the impact of highly hazardous noise in subsequent weighted calculations. The preset noise category detection weight information assigns importance coefficients to the detection results of different noise categories. These coefficients are determined based on factors such as the degree of hazard of the noise type and the detection priority. By multiplying the detection information of each noise category with the corresponding weight coefficient, the detection results are weighted and adjusted to highlight the influence of key noise categories on the final analysis and suppress the interference of secondary noise categories, thereby obtaining weighted information for multiple noise categories.

[0079] Step S504: Based on the multiple noise category detection weighted information and the preset noise category detection analysis bias information, multiple noise detection analysis information is obtained.

[0080] In this embodiment, preset noise category detection and analysis bias information is used to calibrate system errors. Its construction is based on the statistical deviation between historical detection data and actual noise conditions. A correction coefficient compensates for the offset in the detection results. This can be achieved by collecting the difference between the weighted information of noise category detection and the actual noise assessment results from a large amount of historical detection data. For example, statistical analysis of 1000 sets of mechanical vibration noise detection data reveals that the average detection result is 0.2 lower than the actual value; the average detection result of electromagnetic interference noise is 0.1 higher than the actual value; and the detection result of airflow noise is basically consistent with the actual value, with a deviation of 0. Therefore, a bias coefficient is set based on the deviation data to offset systematic deviations. For example, for mechanical vibration noise... The bias coefficient is set to +0.2 to compensate for underestimating detection results, the bias coefficient for electromagnetic interference noise is set to -0.1 to correct overestimating results, and the bias coefficient for airflow noise is set to 0. The bias coefficients are then periodically updated based on new detection data. If the detection deviation for mechanical vibration noise changes from -0.2 to -0.15, the bias coefficient is adjusted from +0.2 to +0.15 to ensure the bias information always adapts to the actual detection environment. This results in the final preset noise category detection and analysis bias information: mechanical vibration noise +0.2, electromagnetic interference noise -0.1, and airflow noise 0. This is used to fine-tune the weighting information in subsequent calculations, making the results closer to the actual noise situation. The preset noise category detection and analysis bias information is a set of constant values ​​used to calibrate the detection results, compensate for system errors, or adjust the detection threshold. Each noise category detection weighting information is added to its corresponding bias value, and the weighting result is fine-tuned to make the final output analysis result more consistent with the characteristics of the actual noise environment. After this adjustment, the weighted information for each noise category is converted into noise detection and analysis information that includes complete information such as noise type, intensity assessment, and potential impact.

[0081] The noise detection and analysis method provided in this application significantly improves the recognition accuracy and classification robustness of complex noise patterns through vector coding, standardization processing, and multi-level weighted analysis. This enables the effective differentiation of the characteristic differences of different types of noise and accurate assessment of their impact, providing a more scientific and targeted decision-making basis for noise control in the manufacturing of silencers and the operating environment of industrial equipment, and further enhancing the practicality and reliability of noise detection and analysis.

[0082] Figure 6 The flowchart illustrating the implementation of the noise detection and analysis method provided in Embodiment Six of this application is shown. The difference between this method and Embodiment Five described above is that:

[0083] Multiple preset noise category detection vectors include preset noise category feature matching detection vectors, preset noise category feature reference detection vectors, and preset noise category feature content detection vectors;

[0084] Step S502 specifically includes:

[0085] Step S601: Calculate multiple noise category feature matching information based on the multiple sound detection category vectors and the preset noise category feature matching detection vectors.

[0086] In this embodiment, the construction of the preset noise category feature matching detection vector requires first collecting a large amount of sample data of known noise categories and extracting key features that represent the noise category from these samples, such as the amplitude of specific frequency bands of mechanical noise and the pulse interval of electromagnetic noise. For example, for a certain type of mechanical operating noise, 1000 samples are selected, and the amplitude features of the three key frequencies of 200Hz, 500Hz, and 800Hz are extracted. The average value of the amplitude at each frequency is calculated, yielding 0.05V, 0.08V, and 0.06V respectively. These average values ​​are then sorted according to the importance of the key features to form the preset noise category feature matching detection vector for this type of noise. The preset noise category feature matching detection vector can be manually set or can be a template vector used to measure the degree of matching between the sound detection category vector and the standard noise features in key dimensions. By comparing the consistency of the corresponding features in each sound detection category vector and the matching detection vector, the overlap ratio of the two in the core features is calculated. The obtained ratio value is the multiple noise category feature matching information, and each piece of information reflects the core matching situation between the corresponding sound detection category and the standard noise features.

[0087] Step S602: Calculate multiple noise category feature reference information based on the multiple sound detection category vectors and the preset noise category feature reference detection vector.

[0088] In this embodiment, the construction of the preset noise category feature reference detection vector involves determining the normal range of various noise types across different feature dimensions. Taking airflow noise as an example, multiple airflow noise samples under normal operating conditions are collected, and the minimum and maximum values ​​of their intensity features are calculated. For instance, the intensity is between 30dB and 60dB, and the frequency is mainly distributed between 100Hz and 1000Hz. This range information is organized according to feature type to form the preset noise category feature reference detection vector, which includes the lower and upper limits of each feature. The preset noise category feature reference detection vector can be manually set or can be a reference template containing typical feature ranges for different noise categories. The feature value in each sound detection category vector is compared with the normal range of the corresponding feature in the reference detection vector to determine whether the feature value is within the range and the degree of deviation. This generates multiple noise category feature reference information, each reflecting the degree of conformity between the corresponding sound detection category feature and the typical noise feature range.

[0089] Step S603: Calculate multiple noise category feature content information based on the multiple sound detection category vectors and the preset noise category feature content detection vectors.

[0090] In this embodiment, the construction of the preset noise category feature content detection vector requires a detailed analysis of the various specific features of noise. For example, for electromagnetic interference noise, its features include pulse frequency, duration, peak voltage, etc. Multiple samples of this type of noise are collected, and the common values ​​of these features are statistically analyzed. For instance, pulse frequency is mostly between 50Hz and 60Hz, duration is mostly between 0.1s and 0.5s, and peak voltage is mostly between 0.2V and 0.5V. These common values ​​and their corresponding feature names are arranged in a certain order to form the preset noise category feature content detection vector. The preset noise category feature content detection vector can be manually set or can be a template used to parse the specific feature content in the sound detection category vector, covering detailed feature descriptions such as the frequency distribution and intensity changes of noise. By comparing each sound detection category vector with this content detection vector, the specific content parameters related to the noise features in the vector are extracted, forming multiple noise category feature content information. Each piece of information contains the specific performance of the corresponding sound detection category in terms of detailed features.

[0091] Step S604: Calculate multiple noise category feature matching detection vectors based on the multiple noise category feature matching information and the multiple noise category feature reference information.

[0092] In this embodiment, the matching information of each noise category feature is comprehensively calculated with the corresponding noise category feature reference information. Combining the degree of matching and range conformity reflected by both, a new vector form of detection result is generated, namely multiple noise category feature matching detection vectors. These vectors integrate the core feature matching situation and range conformity situation, and more comprehensively reflect the overall matching status between the sound detection category and the preset noise category.

[0093] Step S605: Based on the multiple noise category feature matching detection vectors and multiple noise category feature content information, multiple noise category detection information is obtained.

[0094] In this embodiment, the comprehensive matching state contained in each noise category feature matching detection vector is combined with the detailed feature parameters in the corresponding noise category feature content information. By integrating the information of both, a result that can comprehensively reflect the possibility that the sound detection category belongs to a specific noise category and its specific feature performance is formed. These results are multiple noise category detection information.

[0095] The noise detection and analysis method provided in this application improves the precision and accuracy of noise category detection, comprehensively captures the characteristic differences of different noise categories, and makes the generated multiple noise category detection information more targeted and reliable, providing a more accurate basis for subsequent noise detection and analysis, thereby enhancing the depth and effectiveness of the overall noise detection and analysis.

[0096] Corresponding to the method in the above embodiments, Figure 7 A structural block diagram of the noise detection and analysis system provided in the embodiments of this application is shown. For ease of explanation, only the parts related to the embodiments of this application are shown. Figure 7 The noise detection and analysis system in the example can be the execution subject of the noise detection and analysis method provided in the first embodiment above.

[0097] Reference Figure 7 The noise detection and analysis system includes:

[0098] The sound signal acquisition module 710 is used to acquire multiple sound signals to be detected.

[0099] The sound feature information generation module 720 is used to sample, extract features and standardize the multiple sound signals to be detected to obtain multiple sound feature information to be detected.

[0100] The sound detection category information generation module 730 is used to classify the multiple sound feature information to be detected and generate multiple sound detection category information.

[0101] The noise detection and analysis information generation module 740 is used to perform detection and analysis processing on the multiple sound detection category information to generate multiple noise detection and analysis information.

[0102] For details on how each module in the noise detection and analysis system provided in this application implements its respective function, please refer to the foregoing. Figure 1 The description of Embodiment 1 shown will not be repeated here.

[0103] The module 720 for generating sound feature information to be detected specifically includes:

[0104] The unit for generating the sequence information of the sound signal to be detected is used to sample and process the plurality of sound signals to be detected to obtain the sequence information of the plurality of sound signals to be detected;

[0105] The unit for generating subsequence information of the sound signal to be detected is used to segment the multiple subsequence information of the sound signal to be detected based on a preset temporal segmentation interval of the sound signal to be detected, so as to obtain multiple subsequence information of the sound signal to be detected.

[0106] The time-frequency domain feature extraction unit is used to extract time-frequency domain features from the multiple sub-sequence information of the sound signals to be detected, so as to obtain time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected.

[0107] The detection sound feature information generation unit is used to standardize the time-domain feature information and the frequency-domain feature information of the multiple detection sound signals to obtain multiple detection sound feature information.

[0108] For details on how each module in the noise detection and analysis system provided in this application implements its respective function, please refer to the foregoing. Figure 2 The description of Embodiment 2 shown will not be repeated here.

[0109] It should be understood that in this application, the execution order of each step in the embodiments is determined by function and internal logic, and the sequence number does not represent the order of execution and does not constitute a limitation on the implementation process; the term "comprising" does not exclude the existence or addition of other features, steps, etc., and "and / or" refers to any and all possible combinations of related items; "if" and similar phrases should be interpreted according to the context as meaning "when...", "in response to determining / detecting", etc.; terms such as "first" and "second" are only used to distinguish elements and do not imply relative importance, and element names are interchangeable without departing from the scope of the embodiments. In addition, the expressions "one embodiment", "some embodiments", etc. in this application refer to specific features, structures, or characteristics included in one or more embodiments, and such statements appearing in different places do not necessarily refer to the same embodiment, and "comprising", "including", "having", and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0110] The noise detection and analysis method provided in this application can be applied to terminal devices such as mobile phones, tablets, wearable devices, in-vehicle devices, augmented reality (AR) / virtual reality (VR) devices, laptops, ultra-mobile personal computers (UMPCs), netbooks, and personal digital assistants (PDAs). This application does not impose any restrictions on the specific type of terminal device.

[0111] Figure 8 This is a schematic diagram of the structure of a terminal device provided in an embodiment of this application. For example... Figure 8 As shown, the terminal device 8 in this embodiment includes: at least one processor 80 ( Figure 8Only one is shown in the image), and a memory 81 is stored in which a computer program 82 that can run on the processor 80 is stored. When the processor 80 executes the computer program 82, it implements the steps in the various noise detection and analysis method embodiments described above, for example... Figure 1 Steps S101 to S104 are shown. Alternatively, when the processor 80 executes the computer program 82, it implements the functions of each module / unit in the above system embodiments, for example... Figure 7 The functions of modules 710 to 740 are shown.

[0112] The terminal device 8 can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor 80 and a memory 81. Those skilled in the art will understand that... Figure 8 This is merely an example of terminal device 8 and does not constitute a limitation on terminal device 8. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal device may also include input transmission devices, network access devices, buses, etc.

[0113] The processor 80 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0114] In some embodiments, the memory 81 may be an internal storage unit of the terminal device 8, such as a hard disk or memory of the terminal device 8. The memory 81 may also be an external storage device of the terminal device 8, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal device 8. Furthermore, the memory 81 may include both internal and external storage units of the terminal device 8. The memory 81 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 81 can also be used to temporarily store data that has been sent or will be sent.

[0115] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0116] This application also provides a terminal device, which includes at least one memory, at least one processor, and a computer program stored in the at least one memory and executable on the at least one processor. When the processor executes the computer program, it causes the terminal device to implement the steps in any of the above method embodiments.

[0117] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0118] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the various method embodiments above.

[0119] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0120] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0121] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0122] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0123] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A noise detection and analysis method, characterized in that, include: Acquire multiple sound signals to be detected; The multiple sound signals to be detected are sampled, feature extracted, and standardized to obtain multiple sound feature information to be detected. The multiple sound feature information to be detected is classified to generate multiple sound detection category information; The multiple sound detection category information is processed and analyzed to generate multiple noise detection and analysis information. The step of detecting and analyzing the multiple sound detection category information to generate multiple noise detection analysis information specifically includes: The multiple sound detection category information is encoded and normalized to obtain multiple sound detection category vectors; Based on multiple preset noise category detection vectors, the multiple sound detection category vectors are detected and calculated to obtain multiple noise category detection information; Based on the multiple noise category detection information and the preset noise category detection weight information, multiple noise category detection weight information is obtained. The preset noise category detection weight information assigns importance coefficients to the detection results of different noise categories. The coefficients are determined based on the degree of harm of the noise type and the detection priority. Based on the multiple noise category detection weighted information and the preset noise category detection analysis bias information, multiple noise detection analysis information is obtained. The preset noise category detection analysis bias information is a set of constant values ​​used to calibrate the detection results, to compensate for system errors or adjust the detection threshold. Each noise category detection weighted information is added to the corresponding bias value, and the weighted result is fine-tuned so that the final output analysis result is more in line with the characteristics of the actual noise environment. After adjustment, each noise category detection weighted information is converted into noise detection analysis information containing complete content such as noise type, intensity assessment, and potential impact. Multiple preset noise category detection vectors include preset noise category feature matching detection vectors, preset noise category feature reference detection vectors, and preset noise category feature content detection vectors. The construction of the preset noise category feature matching detection vector extracts key features representing noise categories from the samples and serves as a template vector to measure the degree of matching between the sound detection category vector and the standard noise features in key feature dimensions. The construction of the preset noise category feature reference detection vector determines the normal range of various types of noise in different feature dimensions. The construction of the preset noise category feature content detection vector details the specific feature content of various types of noise. The step of detecting and calculating multiple noise category vectors based on multiple preset noise category detection vectors to obtain multiple noise category detection information specifically includes: Based on the multiple sound detection category vectors and the preset noise category feature matching detection vector, the overlap ratio of the two on key features is calculated, and the obtained ratio value is the calculated noise category feature matching information. Based on the multiple sound detection category vectors and the preset noise category feature reference detection vectors, multiple noise category feature reference information is calculated; Based on the multiple sound detection category vectors and the preset noise category feature content detection vectors, multiple noise category feature content information is calculated; Based on the multiple noise category feature matching information and the multiple noise category feature reference information, multiple noise category feature matching detection vectors are calculated; Based on the matching detection vectors of the multiple noise category features and the content information of the multiple noise category features, multiple noise category detection information is obtained.

2. The noise detection and analysis method as described in claim 1, characterized in that, The step of sampling, extracting features, and standardizing the multiple sound signals to be detected to obtain multiple sound feature information specifically includes: The multiple sound signals to be detected are sampled and processed to obtain multiple sound signal sequence information to be detected; Based on the preset temporal segmentation interval of the sound signal to be detected, the multiple sound signal sequence information to be detected is segmented to obtain multiple sound signal sub-sequence information to be detected; Time-frequency domain features are extracted from the multiple sub-sequences of the sound signals to be detected, resulting in time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected. The time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected are standardized to obtain multiple sound feature information to be detected.

3. The noise detection and analysis method as described in claim 1, characterized in that, The step of classifying the multiple sound feature information to be detected and generating multiple sound detection category information specifically includes: Based on the preset number of sound detection categories, the feature information of the multiple sounds to be detected is randomly extracted to obtain multiple core feature information of the sounds to be detected. Calculate the logical distance between the multiple detectable sound feature information and the multiple detectable sound core feature information to obtain the multiple detectable sound feature classification distance information; Based on the classification distance information of the multiple sound features to be detected and the core feature information of the multiple sound features to be detected, the multiple sound feature information to be detected is classified to obtain multiple sound detection category information.

4. The noise detection and analysis method as described in claim 3, characterized in that, The step of classifying the multiple sound feature information to obtain multiple sound detection category information based on the multiple sound feature classification distance information and the multiple sound core feature information to be detected specifically includes: The minimum value of the classification distance information of multiple sound features corresponding to multiple core feature information of the sound to be detected is extracted to obtain the maximum and minimum value information of the classification distance of multiple sound features to be detected. Based on the maximum and minimum values ​​of the classification distances of multiple sound features to be detected and the core feature information of the sound to be detected corresponding to the maximum and minimum values ​​of the classification distances of the sound features to be detected, the sound feature information to be detected corresponding to the maximum and minimum values ​​of the classification distances of multiple sound features to be detected is classified to obtain multiple sound detection feature subclass information. Calculate the mean of multiple sound detection feature subclass information to obtain the core feature information of multiple sound detection subclasses; Determine whether the core feature information of the plurality of sound detection subclasses is the same as the core feature information of the plurality of sounds to be detected; If so, then based on the multiple sound detection feature subclass information, multiple sound detection category information are obtained; If not, the core feature information of the multiple sound detection subclasses is used as the core feature information of multiple sounds to be detected, and the process is returned to the step of calculating the logical distance between the multiple sound feature information to be detected and the multiple sound core feature information to obtain the classification distance information of multiple sounds to be detected.

5. A noise detection and analysis system, characterized in that, include: The sound signal acquisition module is used to acquire multiple sound signals to be detected; The detection sound feature information generation module is used to sample, extract features and standardize the multiple detection sound signals to obtain multiple detection sound feature information; The sound detection category information generation module is used to classify the multiple sound feature information to be detected and generate multiple sound detection category information. The noise detection and analysis information generation module is used to detect, analyze, and process the multiple sound detection category information to generate multiple noise detection and analysis information. The step of detecting and analyzing the multiple sound detection category information to generate multiple noise detection analysis information specifically includes: The multiple sound detection category information is encoded and normalized to obtain multiple sound detection category vectors; Based on multiple preset noise category detection vectors, the multiple sound detection category vectors are detected and calculated to obtain multiple noise category detection information; Based on the multiple noise category detection information and the preset noise category detection weight information, multiple noise category detection weight information is obtained. The preset noise category detection weight information assigns importance coefficients to the detection results of different noise categories. The coefficients are determined based on the degree of harm of the noise type and the detection priority. Based on the multiple noise category detection weighted information and the preset noise category detection analysis bias information, multiple noise detection analysis information is obtained. The preset noise category detection analysis bias information is a set of constant values ​​used to calibrate the detection results, to compensate for system errors or adjust the detection threshold. Each noise category detection weighted information is added to the corresponding bias value, and the weighted result is fine-tuned so that the final output analysis result is more in line with the characteristics of the actual noise environment. After adjustment, each noise category detection weighted information is converted into noise detection analysis information containing complete content such as noise type, intensity assessment, and potential impact. Multiple preset noise category detection vectors include preset noise category feature matching detection vectors, preset noise category feature reference detection vectors, and preset noise category feature content detection vectors. The construction of the preset noise category feature matching detection vector extracts key features representing noise categories from the samples and serves as a template vector to measure the degree of matching between the sound detection category vector and the standard noise features in key feature dimensions. The construction of the preset noise category feature reference detection vector determines the normal range of various types of noise in different feature dimensions. The construction of the preset noise category feature content detection vector details the specific feature content of various types of noise. The step of detecting and calculating multiple noise category vectors based on multiple preset noise category detection vectors to obtain multiple noise category detection information specifically includes: Based on the multiple sound detection category vectors and the preset noise category feature matching detection vector, the overlap ratio of the two on key features is calculated, and the obtained ratio value is the calculated noise category feature matching information. Based on the multiple sound detection category vectors and the preset noise category feature reference detection vectors, multiple noise category feature reference information is calculated; Based on the multiple sound detection category vectors and the preset noise category feature content detection vectors, multiple noise category feature content information is calculated; Based on the multiple noise category feature matching information and the multiple noise category feature reference information, multiple noise category feature matching detection vectors are calculated; Based on the matching detection vectors of the multiple noise category features and the content information of the multiple noise category features, multiple noise category detection information is obtained.

6. The noise detection and analysis system as described in claim 5, characterized in that, The module for generating the sound feature information to be detected specifically includes: The unit for generating the sequence information of the sound signal to be detected is used to sample and process the plurality of sound signals to be detected to obtain the sequence information of the plurality of sound signals to be detected; The unit for generating subsequence information of the sound signal to be detected is used to segment the multiple subsequence information of the sound signal to be detected based on a preset temporal segmentation interval of the sound signal to be detected, so as to obtain multiple subsequence information of the sound signal to be detected. The time-frequency domain feature extraction unit is used to extract time-frequency domain features from the multiple sub-sequence information of the sound signals to be detected, so as to obtain time-domain feature information and frequency-domain feature information of the multiple sound signals to be detected. The detection sound feature information generation unit is used to standardize the time-domain feature information and the frequency-domain feature information of the multiple detection sound signals to obtain multiple detection sound feature information.

7. A terminal device, characterized in that, The terminal device includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 4.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 4.

Images