Noise map updating method and device based on noise monitoring data

Abstract

The invention relates to a noise map updating method and device based on noise monitoring data, and the method comprises the following steps: collecting the noise of a target region through an acoustic sensor, forming an original data packet, carrying out the time-frequency decomposition, and extracting a noise time-frequency feature vector; and carrying out sound source localization based on the feature vector to obtain noise source space distribution data. And calculating a sound ray propagation path, and analyzing a diffraction effect of an obstacle on the path to obtain a diffraction attenuation coefficient. And combining the diffraction attenuation coefficient with the noise source data to carry out sound field superposition, and generating noise distribution superposition data. And finally, incremental updating is performed on the original noise map by using the superposed data, and an updated regional noise map reflecting the real-time noise condition is finally generated, so that the problem of how to efficiently fuse the noise distribution information obtained by real-time analysis with the original noise map to realize low delay is solved.

Description

Technical Field

[0001] The present invention relates to the field of noise technology, and in particular to a noise map updating method and device based on noise monitoring data. Background Art

[0002] With the acceleration of urbanization and the increasing frequency of transportation, industrial and other activities, environmental noise pollution has become a major factor affecting residents' quality of life and public health. Existing methods not only need to fully utilize the high-temporal and spatial resolution noise data obtained by acoustic sensors to extract effective time-frequency features to identify and locate dynamic sound sources, but also need to combine refined sound propagation path analysis and diffraction effect modeling to improve the physical accuracy of sound field prediction. At the same time, how to efficiently integrate the noise distribution information obtained from real-time analysis with the original noise map to achieve low-latency, high-precision incremental map updates is a key technical bottleneck that urgently needs to be overcome in the current field of intelligent sound environment management. Existing technologies still have obvious shortcomings in the fusion of data-driven and physical models, accurate modeling of complex environmental sound propagation, and dynamic map update mechanisms, which limits the real-time application efficiency of noise maps in smart cities and environmental monitoring. Summary of the Invention

[0003] The purpose of the present invention is to solve one of the technical problems existing in the prior art to at least a certain extent.

[0004] To achieve the above objectives, in a first aspect, the present invention provides a noise map updating method based on noise monitoring data, comprising the following steps: The acoustic sensor is used to collect ambient noise in the target area to obtain raw noise data, and the raw noise data is decomposed in the time-frequency domain to obtain a noise time-frequency feature vector; Performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; Calculating a sound ray propagation path of the ambient noise in a target area based on the noise source spatial distribution data, and performing a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; Performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; An incremental update is performed on the preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

[0005] Furthermore, the time-frequency decomposition of the original noise data to obtain the noise time-frequency feature vector includes: Performing short-time Fourier transform on the original noise data to obtain time-spectrogram data, and counting frequency band energy in the time-spectrogram data; Performing multi-scale analysis on the frequency band energy by wavelet packet decomposition to obtain the noise fluctuation coefficient; The noise fluctuation coefficient is subjected to feature dimensionality reduction processing to obtain a noise time-frequency feature vector.

[0006] Furthermore, performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data includes: Extracting the noise characteristic peak value of each channel in the noise time-frequency feature vector, and calculating the time difference of arrival of sound sources between different channels based on the noise characteristic peak value; Constructing a positioning equation group based on the time difference of arrival of the sound source and the spatial coordinates of the acoustic sensor, and iteratively solving the positioning equation group to obtain the coordinates of the sound source; Performing density cluster analysis on the sound source coordinates to obtain sound source clusters, and performing center coordinate calculation on the sound source clusters to obtain cluster center coordinates; The radiation power and influence range of each sound source are calculated based on the cluster center coordinates, and the cluster center coordinates are associated and integrated with the radiation power and influence range to obtain noise source spatial distribution data.

[0007] Furthermore, the calculating of the sound ray propagation path of the ambient noise in the target area based on the noise source spatial distribution data includes: Determining a sound ray emission source based on cluster center coordinates and radiation power in the noise source spatial distribution data, and discretizing the emission direction angle of the sound ray emission source to obtain a sound ray direction parameter; Calling geographic information data of the target area, and performing spatial intersection detection on the geographic information data and the sound line direction parameter to obtain a sound line-geographic action point; The reflection and refraction parameters of the ambient noise in the target area are calculated for the sound line-geographical action point, and the path of the sound line-geographical action point is extended based on the reflection and refraction parameters to obtain the sound line propagation path.

[0008] Furthermore, the performing of a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient includes: Performing spatial tangency detection on the sound ray propagation path and the edge of the obstacle in the target area to obtain a diffraction critical point set, and performing Fresnel zone occlusion analysis on the diffraction critical point set to obtain a Fresnel zone parameter value; Calculating a diffraction coefficient base value based on the Fresnel zone parameter value, and performing energy attenuation conversion on the diffraction coefficient base value and the diffraction critical point set to obtain a diffraction energy loss value; The diffraction effect of the ambient noise in the target area is calculated based on the diffraction energy loss value to obtain a diffraction attenuation coefficient.

[0009] Furthermore, performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data includes: Performing sound pressure level conversion on the radiation power in the noise source spatial distribution data to obtain sound source sound pressure level data, and performing spatial correlation matching on the sound source sound pressure level data and the diffraction attenuation coefficient to obtain sound source-attenuation correlation data; Performing a sound pressure level superposition calculation based on the sound source-attenuation correlation data to obtain a preliminary superposition sound pressure level, and performing environmental correction on the preliminary superposition sound pressure level to obtain a corrected superposition sound pressure level; The time characteristics of the corrected superimposed sound pressure levels are integrated to obtain time-weighted superimposed data, and the time-weighted superimposed data are formatted for spatial distribution to obtain noise distribution superimposed data.

[0010] In a second aspect, the present invention further provides a noise map updating device based on noise monitoring data, comprising: An acquisition module is used to collect ambient noise in a target area through an acoustic sensor to obtain raw noise data, and to perform time-frequency domain decomposition on the raw noise data to obtain a noise time-frequency feature vector; An analysis module is used to perform sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; a calculation module, configured to calculate a sound ray propagation path of the ambient noise in a target area based on the noise source spatial distribution data, and perform a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; a superposition module, configured to perform sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; An updating module is used to incrementally update a preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

[0011] In a third aspect, an embodiment of the present invention provides a noise map updating device based on noise monitoring data, comprising: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor is enabled to implement the noise map updating method based on noise monitoring data.

[0012] In a fourth aspect, an embodiment of the present invention provides a storage medium storing a program executable by a processor, wherein the program executable by the processor is used to implement the noise map updating method based on noise monitoring data when executed by the processor.

[0013] The present invention provides a noise map updating method based on noise monitoring data, comprising the following steps: collecting ambient noise in a target area through an acoustic sensor to obtain raw noise data, and performing time-frequency decomposition on the raw noise data to obtain a noise time-frequency feature vector; performing sound source positioning analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; calculating a sound ray propagation path of the ambient noise in the target area based on the noise source spatial distribution data, and performing diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; and incrementally updating a preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map. This method solves the problem of how to efficiently fuse the noise distribution information obtained by real-time analysis with the original noise map to achieve low latency, avoids the high computational cost of full recalculation, and only performs local correction and fusion on the changed area, significantly improving map update efficiency and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 1 is a schematic diagram of the steps of a noise map updating method based on noise monitoring data in one embodiment of the present invention; Figure 2 is a structural block diagram of a noise map updating device based on noise monitoring data in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0015] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0016] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and are not to be construed as limiting the present invention. The step numbers in the following embodiments are provided for ease of explanation only and do not limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0017] A noise map updating method based on noise monitoring data and an apparatus according to an embodiment of the present invention will be described in detail below with reference to the accompanying drawings. First, a noise map updating method based on noise monitoring data according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0018] like Figure 1 As shown, Figure 1 A noise map updating method based on noise monitoring data in one embodiment of the present invention includes the following steps: Step S1: collect environmental noise in a target area through an acoustic sensor to obtain original noise data, and perform time-frequency domain decomposition on the original noise data to obtain a noise time-frequency feature vector.

[0019] Specifically, the environmental noise of the target area is collected by acoustic sensors to obtain raw noise data, and the raw noise data is decomposed in the time-frequency domain to obtain the noise time-frequency feature vector. This process first relies on the distributed acoustic sensor network deployed in the target area to continuously collect sound signals in the environment, thereby forming raw noise data in the form of time series. These data contain information on the sound pressure changes of various sound sources at different times. Then, in order to reveal the dynamic characteristics of the noise signal in the two dimensions of time and frequency, the raw noise data needs to be decomposed in the time-frequency domain. This processing usually uses mathematical tools such as short-time Fourier transform (STFT), wavelet transform or Hilbert-Huang transform to convert the one-dimensional time domain signal into a two-dimensional time-frequency representation matrix, so that the noise can be clearly identified. The distribution of energy in different frequency components and its evolution over time are analyzed to extract recognizable noise time-frequency feature vectors. These feature vectors not only contain the energy intensity of each frequency band, but also retain key information such as the occurrence time, duration and frequency evolution trajectory of the sound source event. Therefore, they can effectively distinguish the broadband characteristics of traffic noise from the periodic pulse characteristics of construction machinery. For example, in the application scenario of urban road intersections, the original noise data collected by the acoustic sensor may be a mixture of vehicle horns, engine roars and tire rolling noises. By decomposing these data in the time-frequency domain, the concentrated energy clusters of high-frequency horn signals in specific time periods can be separated, thereby forming noise time-frequency feature vectors with spatial and temporal directionality, providing an accurate data basis for subsequent sound source localization analysis.

[0020] Step S2: performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data.

[0021] Specifically, based on the noise time-frequency feature vector, the target area is subjected to sound source localization analysis to obtain noise source spatial distribution data. This process is based on obtaining the noise time-frequency feature vector generated by time-frequency domain decomposition of the original noise data. The spatial geometric arrangement relationship of multiple acoustic sensors and the differences in time delay, phase difference and energy distribution of the noise signals collected by them are utilized, combined with array signal processing or sound source imaging algorithm, to perform spatial inversion calculation on the key information representing the dynamic characteristics of the sound source in the noise time-frequency feature vector, thereby determining the spatial position and intensity distribution of each noise source in the target area. In specific implementation, beamforming, matched field processing or sparse signal reconstruction and other methods can be used. The sound propagation paths in different directions are scanned and matched in the frequency domain or time-frequency domain. By comparing the signal coherence or energy focusing degree at each candidate position, the most likely sound source coordinates are identified, and then the noise source spatial distribution data including the sound source location, type and relative intensity are generated. For example, in the application scenario of urban road intersections, when multiple vehicles are driving at the same time and generating overlapping noise, by jointly analyzing the noise time-frequency feature vectors obtained by acoustic sensors distributed around the intersection, it is possible to distinguish engine noise or brake sounds from different lanes and accurately locate their spatial positions, so that the noise source spatial distribution data can truly reflect the dynamic distribution of mobile sound sources in the traffic flow, providing accurate input basis for subsequent sound propagation path calculations and map updates.

[0022] Step S3: calculating the sound ray propagation path of the ambient noise in the target area based on the noise source spatial distribution data, and performing diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient.

[0023] Specifically, the sound ray propagation path of the ambient noise in the target area is calculated based on the spatial distribution data of the noise source, and the diffraction effect of the obstacles in the target area is analyzed based on the sound ray propagation path to obtain the diffraction attenuation coefficient. This step first uses the position coordinates and initial sound energy information of each sound source contained in the obtained noise source spatial distribution data, combined with the geographic information system (GIS) data or three-dimensional building model of the target area, and adopts a sound ray tracing algorithm to simulate the path of ambient noise propagation from each sound source point to the surrounding space. These sound ray propagation paths include not only direct paths of straight-line propagation, but also reflection paths occurring on the surfaces of obstacles such as building walls and the ground, as well as diffraction paths occurring at the edges or tops of buildings. When the sound ray propagation path interacts with the obstacles in the target area during the propagation process, When sound waves pass through the edge or top of an obstacle, further diffraction effect analysis is required, especially when the sound waves pass through the edge or top of an obstacle. By applying Keller's geometric theory of diffraction (GTD) or its modified form, the unified theory of diffraction (UTD), the additional attenuation caused by energy scattering when the sound waves bypass the obstacle is calculated, thereby accurately obtaining the diffraction attenuation coefficient corresponding to each affected sound line. For example, in the application scenario of an urban road intersection, when the noise generated by a certain vehicle propagates from the main road to the secondary sidewalk area blocked by high-rise buildings, its sound line propagation path will pass through the corner of the building and be diffracted. At this time, based on the spatial geometric relationship between this path and the building, combined with the sound absorption characteristics of the material, the diffraction attenuation coefficient under this path can be calculated. This coefficient will be used in the subsequent sound field superposition process to improve the accuracy of noise distribution prediction.

[0024] Step S4: performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data.

[0025] Specifically, the diffraction attenuation coefficient and the noise source spatial distribution data are superimposed on the sound field to obtain noise distribution superposition data. This process is based on the obtained noise source spatial distribution data and the diffraction attenuation coefficient corresponding to each sound line propagation path, and the sound pressure contribution generated by each noise source at each receiving point in the target area is physically modeled and energy synthesized. Specifically, first, according to the position and initial sound power of each sound source in the noise source spatial distribution data, combined with the geometric divergence attenuation of the sound wave in the free field, the diffraction attenuation coefficient calculated by the above steps is introduced to correct the energy loss in the propagation path, thereby obtaining the total sound energy attenuation from each sound source to the target grid point, and then in the discretized grid space of the target area, the sound pressure levels generated by all sound sources at this point are logarithmically superimposed (that is, the logarithm is taken after the energy is added). ), comprehensively considering the intensity, spatial position, propagation path and diffraction effects caused by obstacles of each sound source, and ultimately generating noise distribution superposition data that reflects the actual sound field distribution characteristics. This data not only reflects the strength distribution of the sound source, but also includes the spatial modulation effect of the complex urban environment on sound propagation. For example, in the application scenario of urban road intersections, when multiple vehicles are running in different lanes, their noise is blocked by surrounding buildings during the propagation process and diffracted. At this time, by combining the diffraction attenuation coefficient corresponding to each vehicle source with its distribution position in space, the sound field energy is superimposed point by point, which can accurately restore the actual noise level in the sidewalk or back street area. In particular, it shows higher prediction accuracy in areas such as acoustic shadows that are easily underestimated by traditional models, thereby providing highly reliable data support for subsequent incremental updates to the original noise map.

[0026] Step S5: incrementally updating the preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

[0027] Specifically, the preset original noise map is incrementally updated based on the noise distribution overlay data to obtain an updated regional noise map. This step is based on the generated noise distribution overlay data, and it is dynamically fused and locally corrected with the pre-stored original noise map at the spatial grid level. The original noise map is usually constructed based on historical data or static models, covering the background noise distribution of the entire target area, while the noise distribution overlay data reflects the latest noise field changes currently collected by the acoustic sensor in real time and obtained after sound source positioning, propagation path analysis and diffraction effect compensation. After the two are aligned through spatial registration, difference detection and weight fusion are performed on areas with significant differences with the support of the geographic information system (GIS). , only the local grids showing obvious sound level changes in the noise distribution overlay data are updated, rather than redrawing the entire map, thereby achieving efficient and low-computational-cost incremental updates. For example, in the application scenario of urban road intersections, when sudden traffic congestion or temporary construction causes a sharp increase in noise levels within a certain period of time, the noise distribution overlay data will capture the significant increase in sound pressure levels in the abnormal area. Based on this data, the system will locally refresh the corresponding road section in the original noise map and the surrounding pedestrian areas affected by diffraction, retaining the map information of the unchanged areas and only updating the noise values ​​of the affected grids. Finally, an updated regional noise map is generated that can reflect traffic dynamics and environmental changes in real time, ensuring the timeliness and accuracy of the map.

[0028] In a specific embodiment, performing time-frequency domain decomposition on the original noise data to obtain a noise time-frequency feature vector includes: Performing short-time Fourier transform on the original noise data to obtain time-spectrogram data, and counting frequency band energy in the time-spectrogram data; Performing multi-scale analysis on the frequency band energy by wavelet packet decomposition to obtain the noise fluctuation coefficient; The noise fluctuation coefficient is subjected to feature dimensionality reduction processing to obtain a noise time-frequency feature vector.

[0029] Specifically, the original noise data is decomposed in the time-frequency domain to obtain a noise time-frequency feature vector. This process specifically includes performing a short-time Fourier transform on the original noise data to obtain time-frequency spectrum data, and counting the frequency band energy in the time-frequency spectrum data; performing multi-scale analysis on the frequency band energy through wavelet packet decomposition to obtain a noise fluctuation coefficient; and performing feature dimensionality reduction processing on the noise fluctuation coefficient to obtain a noise time-frequency feature vector. This series of operations aims to extract recognizable and representative feature information from complex environmental noise signals to support subsequent sound source identification and positioning analysis. First, after obtaining the original noise data by collecting the environmental noise of the target area through the acoustic sensor, since the noise signal is usually a non-stationary random signal, its frequency component changes continuously with time, it is difficult to fully describe its dynamic characteristics by directly analyzing it in the time domain or frequency domain. Therefore, it is necessary to use short-time Fourier transform (STFT) to process the original noise data, that is, to divide the signal into multiple overlapping time windows, and approximately regard it as a stationary signal in each time window and perform Fourier transform, thereby generating a two-dimensional time-spectrum diagram data, which can intuitively show the different frequency components in time. The energy distribution on the axis, for example, in the application scenario of urban road intersections, the low-frequency roar when the vehicle starts, the medium- and high-frequency friction sound when braking, and the short high-frequency pulses produced by honking will all appear as energy concentration in a specific time-frequency area in the time-spectrogram data. Then, the frequency band energy in the time-spectrogram data is further counted, that is, the entire frequency range is divided into several sub-bands (such as 31.5Hz, 63Hz, 125Hz and other octaves), and the energy integral value of each frequency band in different time periods is calculated to form a frequency band energy sequence, which is used to quantify the intensity change trend of the noise in each frequency band. Subsequently, in order to deeply explore the local fluctuation characteristics of frequency band energy in time and scale, especially for the identification needs of sudden, transient or periodic noise events, wavelet packet decomposition is introduced to perform multi-scale analysis of the frequency band energy. Compared with traditional wavelet transform, wavelet packet decomposition has higher time-frequency resolution and finer frequency band division capabilities. It can recursively decompose the low-frequency and high-frequency parts at the same time, thereby capturing the fluctuation details of noise energy at different time scales, and output a set of noise fluctuation coefficients reflecting the energy fluctuation amplitude and frequency response characteristics. These coefficients can effectively characterize the time-frequency patterns of dynamic behaviors such as traffic flow pulsation, vehicle acceleration / deceleration process or intermittent construction noise.Finally, since the noise fluctuation coefficients generated by wavelet packet decomposition have high dimensions and may contain redundant information, it is necessary to perform feature dimensionality reduction processing on them. Methods such as principal component analysis (PCA) or linear discriminant analysis (LDA) are usually used to compress data dimensions while retaining the main information, eliminate the correlation between variables, and ultimately obtain a compact and representative noise time-frequency feature vector. This vector not only contains the frequency band energy distribution characteristics of the original noise signal, but also integrates its dynamic fluctuation characteristics at multiple scales, providing a high-quality, low-redundancy, and high-discrimination input data foundation for the subsequent sound source localization analysis based on the noise time-frequency feature vector. In actual applications at urban road intersections, it is precisely relying on this series of sophisticated time-frequency analysis processes that the system can accurately distinguish between the idling noise of buses and the high-frequency noise generated by the rapid passage of electric bicycles, and convert it into effective feature information that can be used for spatial positioning, thereby ensuring the accuracy and real-time performance of the entire noise map update method.

[0030] In a specific embodiment, performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data includes: Extracting the noise characteristic peak value of each channel in the noise time-frequency feature vector, and calculating the time difference of arrival of sound sources between different channels based on the noise characteristic peak value; Constructing a positioning equation group based on the time difference of arrival of the sound source and the spatial coordinates of the acoustic sensor, and iteratively solving the positioning equation group to obtain the coordinates of the sound source; Performing density cluster analysis on the sound source coordinates to obtain sound source clusters, and performing center coordinate calculation on the sound source clusters to obtain cluster center coordinates; The radiation power and influence range of each sound source are calculated based on the cluster center coordinates, and the cluster center coordinates are associated and integrated with the radiation power and influence range to obtain noise source spatial distribution data.

[0031] Specifically, based on the noise time-frequency feature vector, the target area is subjected to sound source positioning analysis to obtain noise source spatial distribution data. The process specifically includes extracting the noise characteristic peak value of each channel in the noise time-frequency feature vector, and calculating the sound source arrival time difference between different channels based on the noise characteristic peak value; constructing a positioning equation group based on the sound source arrival time difference combined with the spatial coordinates of the acoustic sensor, and iteratively solving the positioning equation group to obtain the sound source coordinates; performing density clustering analysis on the sound source coordinates to obtain sound source clusters, and calculating the center coordinates of the sound source clusters to obtain cluster center coordinates; calculating the radiation power and influence range of each sound source based on the cluster center coordinates, and correlating and integrating the cluster center coordinates with the radiation power and influence range to obtain noise source spatial distribution data. This series of operations aims to accurately restore the distribution state of the sound source in space from the noise signals collected by multiple sensors. First, after obtaining the noise time-frequency feature vector generated by the previous steps, since this vector contains the joint feature information of multiple acoustic sensor channels in the time-frequency domain, the noise characteristic peaks of each channel in the noise time-frequency feature vector can be extracted by analyzing the noise energy response of each channel in the same or related time-frequency units. These peaks generally correspond to the strongest response time and frequency components of significant acoustic events at different sensor locations. For example, in the application scenario of an urban road intersection, when a heavy truck passes, its engine low-frequency noise will sequentially generate energy peaks on multiple acoustic sensors distributed at the four corners of the intersection. These peaks have slight differences in time, which are the arrival time differences caused by the different sound wave propagation paths. Next, based on the differences in the occurrence time of the noise characteristic peaks between each channel, the time difference of arrival (TDOA) of the sound sources between different channels is calculated. This time difference reflects the relative difference in the time required for sound waves to propagate from the same sound source to different sensors and is a key physical quantity for achieving spatial positioning. Subsequently, a hyperbolic localization model based on TDOA is established using the known spatial coordinates of the acoustic sensors and the sound velocity constant. This involves constructing a set of localization equations, each representing a hyperbolic constraint determined by the time difference of arrival between the two sensors. Multiple equations are combined to form a nonlinear system of equations. This system of localization equations is then iteratively solved using numerical optimization methods such as least squares or Newton's method, gradually approximating the true sound source location and obtaining a preliminary set of sound source coordinates. Due to the presence of multiple sound sources, reflected noise, and measurement errors in real environments, direct solution can produce a large number of discrete and redundant candidate points. Therefore, further density clustering analysis is required on the sound source coordinates. Using a spatial density-based clustering algorithm such as DBSCAN, regions with densely distributed coordinate points are identified, forming multiple sound source clusters. Each cluster represents a potential true sound source activity area. The center coordinates of each cluster are then calculated, such as by taking the geometric centroid or weighted average position of all points within the cluster, to obtain a more representative cluster center coordinate, which serves as the final localization result for the sound source.Finally, based on the cluster center coordinates, combined with the sound propagation model and the sound pressure level data measured by the sensor, the radiated power of each sound source is inferred, and its impact range is estimated based on the environmental attenuation characteristics. The cluster center coordinates are then associated and integrated with the radiated power and impact range to form a complete description of the location, intensity, and spatial scope. The final output is the spatial distribution data of the noise source, providing accurate input for subsequent sound propagation path calculations and map updates. For example, in an intersection scenario, the system can identify a strong noise cluster formed when a bus stops and determine that its cluster center is located near the platform. At the same time, its radiated power is calculated to be as high as 85dB(A), and its impact range covers the sidewalk within 20 meters, thereby accurately constructing the spatial distribution information of this dynamic sound source.

[0032] In a specific embodiment, constructing a positioning equation group based on the time difference of arrival of the sound source in combination with the spatial coordinates of the acoustic sensor includes: Performing sensor pair association processing on the time difference of arrival of the sound source to obtain a time difference association set, and performing corresponding sensor pair coordinate extraction on the spatial coordinates of the acoustic sensor to obtain a coordinate pair set; Constructing a hyperbola equation based on the time difference association set and the coordinate pair set to obtain an initial positioning equation set, and performing a geometric configuration solvability analysis on the initial positioning equation set to obtain a solvability index; Performing equation screening on the initial positioning equation set based on the solvability index to obtain a valid equation set, and normalizing the valid equation set to obtain a normalized equation set; A matrix condition number evaluation is performed on the normalized equation set to obtain a condition value, and a weighted adjustment is performed on the normalized equation set based on the condition value to obtain a positioning equation group.

[0033] Specifically, based on the time difference of arrival of the sound source and the spatial coordinates of the acoustic sensor, a positioning equation group is constructed. The process specifically includes performing sensor pair association processing on the time difference of arrival of the sound source to obtain a time difference association set, and extracting the corresponding sensor pair coordinates of the spatial coordinates of the acoustic sensor to obtain a coordinate pair set; constructing hyperbolic equations based on the time difference association set and the coordinate pair set to obtain an initial positioning equation set, and performing geometric configuration solvability analysis on the initial positioning equation set to obtain a solvability index; based on the solvability index, the initial positioning equation set is screened to obtain a valid equation set, and the valid equation set is normalized to obtain a normalized equation set; performing matrix condition number evaluation on the normalized equation set to obtain a conditional value, and performing weighted adjustment on the normalized equation set based on the conditional value to obtain a positioning equation group. This series of operations is intended to improve the accuracy and robustness of sound source positioning. First, after obtaining the time difference of arrival of sound sources between different channels, it is necessary to clarify the two acoustic sensors corresponding to each pair of time difference data. Therefore, the sound source arrival time difference must be subjected to sensor pair association processing, that is, a mapping relationship between the time difference value and the sensor pair it generates is established to form a structured time difference association set. For example, in the application scenario of an urban road intersection, if sensors A, B, C, and D are located at the four corners of the intersection, then any two-by-two combinations between A and B, A and C, etc. constitute multiple sensor pairs, and each sensor pair corresponds to an arrival time difference value, thereby obtaining a complete time difference association set; at the same time, the spatial coordinates of each sensor are extracted from the system's preset deployment information, and the corresponding sensor pair coordinates are extracted according to the above-mentioned sensor pair combination method to form a matching coordinate pair set to ensure that each time difference data is supported by an accurate spatial geometric relationship. Next, using the time difference and coordinate information of each sensor pair, and based on the assumption of constant speed of sound waves propagating in air, a hyperbolic equation describing the sound source position is constructed. This is because the distance difference between the sound source and the two sensors is equal to the speed of sound multiplied by the time difference of arrival. This converts the time difference-coordinate pair of each sensor pair into a nonlinear hyperbolic equation. All equations together constitute the initial positioning equation set. However, due to unreasonable sensor layout in actual deployment (such as collinearity or close spacing), some equations may have a small contribution to positioning or even introduce errors. Therefore, the initial positioning equation set needs to be subjected to geometric configuration solvability analysis. By calculating the spatial distribution angle, baseline length, and relative orientation of each sensor pair, the cross-correlation characteristics of the hyperbola constructed are evaluated, thereby quantifying the solvability index, which reflects the reliability and information gain of the equation in the joint solution.Subsequently, the initial positioning equation set is screened based on the solvability index, equations with low solvability, redundancy or pathological conditions are eliminated, and effective equations with strong positioning capabilities are retained to form an effective equation set; in order to further eliminate the numerical instability problem caused by the dimension or order of magnitude difference of different equations, the effective equation set is normalized, for example, the coefficients of each equation are divided by its modulus or maximum value, so that all equations are on a uniform scale, and a normalized equation set is obtained. Finally, in order to optimize the numerical solution performance of the equation set, the matrix condition number of the normalized equation set is evaluated, and the condition value of its coefficient matrix is ​​calculated. The larger the condition value, the closer the equation set is to singularity and the more unstable the solution. Therefore, based on the condition value, the normalized equation set is weighted and adjusted, and equations that are sensitive to conditions are given lower weights, while equations with high stability are retained with higher weights, thereby balancing the contributions of each equation, and finally forming a positioning equation set that is numerically stable, geometrically reasonable, and has strong anti-interference ability, providing a high-quality mathematical foundation for subsequent iterative solutions. For example, in an intersection scenario, when a construction vehicle starts at the northwest corner, the system constructs an effective hyperbolic equation by screening sensor pairs with good spatial angles such as southeast-southwest and northeast-northwest, and then accurately infers the location of the sound source at the road corner after weighted optimization, significantly improving positioning accuracy.

[0034] In a specific embodiment, the calculating the sound ray propagation path of the ambient noise in the target area based on the noise source spatial distribution data includes: Determining a sound ray emission source based on cluster center coordinates and radiation power in the noise source spatial distribution data, and discretizing the emission direction angle of the sound ray emission source to obtain a sound ray direction parameter; Calling geographic information data of the target area, and performing spatial intersection detection on the geographic information data and the sound line direction parameter to obtain a sound line-geographic action point; The reflection and refraction parameters of the ambient noise in the target area are calculated for the sound line-geographical action point, and the path of the sound line-geographical action point is extended based on the reflection and refraction parameters to obtain the sound line propagation path.

[0035] Specifically, the sound line propagation path of the ambient noise in the target area is calculated based on the spatial distribution data of the noise source. This process specifically includes determining the sound line emission source based on the cluster center coordinates and radiation power in the spatial distribution data of the noise source, and discretizing the emission direction angle of the sound line emission source to obtain the sound line direction parameter; calling the geographic information data of the target area, and performing spatial intersection detection on the geographic information data and the sound line direction parameter to obtain the sound line-geographic action point; calculating the reflection and refraction parameters of the ambient noise in the target area for the sound line-geographic action point, and extending the path of the sound line-geographic action point based on the reflection and refraction parameters to obtain the sound line propagation path. This series of operations is intended to accurately simulate the propagation behavior of noise in a complex urban environment. First, given the spatial distribution data of noise sources, which includes the cluster center coordinates and their corresponding radiation power obtained through cluster analysis, these cluster center coordinates represent the spatial location of the actual noise source, such as frequently occurring traffic concentration points or construction machinery operation points on the road, while the radiation power reflects the intensity level of the sound source. Therefore, each cluster center coordinate can be used as the starting position of sound wave emission, that is, the sound ray emission source, to initiate the sound ray tracing process. In order to fully cover the possible directions of radiation from the sound source to the surrounding space, the emission direction angle of the sound ray emission source needs to be discretized. That is, the omnidirectional space centered on the sound source (usually a horizontal plane from 0° to 360° or a three-dimensional space including elevation angles) is divided into multiple equally spaced direction angles. For example, an emission direction is set every 5° or 10° to form a set of discretized sound ray direction parameters. Each direction parameter corresponds to the propagation direction vector of an initial sound ray, thereby ensuring that the sound ray can simulate the diffusion trend of sound energy in all directions. Next, the system uses geographic data for the target area, typically including three-dimensional spatial structural information such as building outlines, heights, road layouts, green belts, and terrain elevations, to construct a geometric environmental model for sound propagation. Each sound ray, with its directional parameters, is then extended from the source along a specified direction and spatially intersected with obstacle surfaces in the geographic data to determine whether it intersects with building walls, roof edges, or other terrain features along the propagation path. If so, the intersection is recorded, representing the sound ray-geographic point of impact, marking the location where the sound ray first interacts with a physical obstacle. For each sound ray-geographic point of impact, the system then calculates the reflection and refraction parameters of the ambient noise in the target area. Reflection parameters include the reflection angle, reflection coefficient (dependent on the material's sound absorption properties), and energy attenuation, while refraction parameters involve the change in propagation direction and transmission loss when sound waves pass through different media (such as glass curtain walls or vegetation layers). The calculation of these parameters relies on a database of material acoustic properties and physical acoustic models, such as reflection theory based on impedance boundary conditions or Snell's law for refraction analysis.On this basis, the path of the sound ray-geographical action point is extended based on the reflection and refraction parameters. That is, a new sound ray segment is generated according to the propagation direction after reflection or refraction, and it is used as the input for the next round of spatial intersection detection to achieve multi-level propagation tracking of the sound ray until the sound ray energy falls below the threshold or exceeds the target area. Finally, a complete set of sound ray propagation paths is formed, including the direct path, single / multiple reflection paths, and pre-diffraction paths. For example, in the application scenario of an urban road intersection, when a heavy truck starts at the southeast corner, its cluster center coordinates are determined as the sound ray emission source. The system emits sound rays along multiple discrete directions. During the propagation process, one of the sound rays intersects with the glass curtain wall of the opposite high-rise building, forming a sound ray-geographical action point. After calculation, this point has a high reflection coefficient and a certain amount of energy loss. The sound ray is reflected to the sidewalk area of ​​the adjacent street, thereby accurately simulating the phenomenon of noise pollution caused by building reflection to the originally non-direct view area, providing refined propagation path support for subsequent diffraction effect analysis and sound field superposition.

[0036] In a specific embodiment, performing a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient includes: Performing spatial tangency detection on the sound ray propagation path and the edge of the obstacle in the target area to obtain a diffraction critical point set, and performing Fresnel zone occlusion analysis on the diffraction critical point set to obtain a Fresnel zone parameter value; Calculating a diffraction coefficient base value based on the Fresnel zone parameter value, and performing energy attenuation conversion on the diffraction coefficient base value and the diffraction critical point set to obtain a diffraction energy loss value; The diffraction effect of the ambient noise in the target area is calculated based on the diffraction energy loss value to obtain a diffraction attenuation coefficient.

[0037] Specifically, based on the sound ray propagation path, the diffraction effect of the obstacles in the target area is analyzed to obtain the diffraction attenuation coefficient. The process specifically includes performing spatial tangent detection on the edge of the sound ray propagation path and the obstacle in the target area to obtain a diffraction critical point set, and performing Fresnel zone occlusion analysis on the diffraction critical point set to obtain a Fresnel zone parameter value; calculating the diffraction coefficient base value based on the Fresnel zone parameter value, and performing energy attenuation conversion between the diffraction coefficient base value and the diffraction critical point set to obtain a diffraction energy loss value; calculating the diffraction effect of the ambient noise in the target area based on the diffraction energy loss value to obtain the diffraction attenuation coefficient. This series of steps aims to accurately quantify the energy attenuation caused by the diffraction phenomenon when the sound wave encounters the edge of the obstacle, thereby improving the physical reality of the noise distribution prediction. First, based on the generated sound ray propagation path, it is necessary to identify key locations that may cause diffraction. Therefore, spatial tangency detection is performed on the geometric edges of the sound ray propagation path and obstacles in the target area (such as building exterior walls, roof edges, sound barriers, etc.). In other words, it is determined whether the sound ray passes over the edge or corner of the obstacle in a manner close to a tangent. If specific spatial proximity and direction angle conditions are met, the contact point is recorded as a diffraction critical point. All such points constitute a diffraction critical point set, which represents the spatial locations where the sound wave is most likely to undergo significant diffraction behavior. For example, in the application scenario of an urban road intersection, when a sound ray emitted from a main road approaches the vertical corner edge of a high-rise building during propagation, even if no direct collision or reflection occurs, as long as its path forms a small angle with the edge and is within a certain distance range, the system will determine it as a potential diffraction point and include it in the diffraction critical point set. Subsequently, a Fresnel zone obstruction analysis is performed for each critical diffraction point. The Fresnel zone is a spatial model that describes the region in the sound wave propagation path that allows waves to bypass obstacles. It is usually divided into multiple concentric ellipsoidal zones, and its boundaries are determined by the geometric relationship between the sound source, the receiving point, and the obstacle. By calculating the extra path difference of the sound line path relative to the ideal straight path and combining it with the sound wave frequency, the Fresnel zone number and corresponding Fresnel zone parameter value of the diffraction point can be determined. This parameter value reflects the degree of obstruction of the direct sound path by the obstacle and the strength of the diffraction effect. Next, based on the Fresnel zone parameter value, the base value of the diffraction coefficient at that point is calculated using empirical formulas or lookup tables in classical diffraction theory (such as Keller's geometric diffraction theory (GTD) or its modified form (UTD)). This coefficient represents the complex amplitude attenuation and phase change when the sound wave bypasses this edge under ideal conditions. Then, the diffraction coefficient base value is combined with the spatial position information of the diffraction critical point concentration to perform energy attenuation conversion, that is, the square of the modulus of the complex coefficient is converted into the sound intensity attenuation ratio, and correction is made considering factors such as propagation distance, material edge roughness and atmospheric absorption, so as to obtain the diffraction energy loss value of each sound line after passing through the diffraction point.Finally, the energy loss values ​​along all diffraction paths are combined to assess the overall sound energy attenuation of ambient noise in the target area due to diffraction effects. This leads to the diffraction attenuation coefficient, measured in decibels (dB), which is used to compensate for the sound pressure level of the affected sound rays during the subsequent sound field superposition process. For example, on the sidewalk north of the intersection, although blocked by an office building and in the sound shadow zone, the sound rays of traffic noise from the southbound main road form multiple diffraction critical points at the southeast corner edge of the building. After Fresnel zone analysis and energy loss calculation, a significant diffraction attenuation coefficient (e.g., 15dB) is obtained, indicating that considerable noise energy is still propagating into this area through edge diffraction. Based on this, the system retains the contribution of this path in the noise distribution superposition, thus avoiding the serious underestimation of the noise level in this area by traditional models and ensuring that the updated regional noise map more realistically reflects the sound field distribution characteristics in complex urban environments.

[0038] In a specific embodiment, performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data includes: Performing sound pressure level conversion on the radiation power in the noise source spatial distribution data to obtain sound source sound pressure level data, and performing spatial correlation matching on the sound source sound pressure level data and the diffraction attenuation coefficient to obtain sound source-attenuation correlation data; Performing a sound pressure level superposition calculation based on the sound source-attenuation correlation data to obtain a preliminary superposition sound pressure level, and performing environmental correction on the preliminary superposition sound pressure level to obtain a corrected superposition sound pressure level; The time characteristics of the corrected superimposed sound pressure levels are integrated to obtain time-weighted superimposed data, and the time-weighted superimposed data are formatted for spatial distribution to obtain noise distribution superimposed data.

[0039] Specifically, the diffraction attenuation coefficient and the noise source spatial distribution data are superimposed on the sound field to obtain noise distribution superposition data. The process specifically includes converting the radiation power in the noise source spatial distribution data into sound pressure level to obtain sound source sound pressure level data, and spatially correlating the sound source sound pressure level data with the diffraction attenuation coefficient to obtain sound source-attenuation correlation data; performing sound pressure level superposition calculation based on the sound source-attenuation correlation data to obtain a preliminary superimposed sound pressure level, and performing environmental correction on the preliminary superimposed sound pressure level to obtain a corrected superimposed sound pressure level; integrating the time characteristics of the corrected superimposed sound pressure level to obtain time-weighted superimposed data, and spatially formatting the time-weighted superimposed data to obtain noise distribution superimposed data. This series of operations aims to deeply integrate the propagation characteristics of physical modeling with the measured sound source information to generate high-precision dynamic sound field distribution. First, under the premise of knowing the spatial distribution data of the noise source, which includes the cluster center coordinates of each sound source and its corresponding radiation power, the physical power quantity needs to be converted into a sound pressure level form that can participate in the sound field calculation. Therefore, the radiation power is converted into a sound pressure level. According to the free field point sound source propagation model and combined with the distance from the sound source to the target grid point, the theoretical sound pressure level generated at each spatial position is calculated to obtain the sound source sound pressure level data covering the target area; at the same time, the diffraction attenuation coefficient obtained in the above steps reflects the additional energy loss caused by the diffraction of the obstacle edge during the propagation of the sound wave. These attenuation coefficients need to be accurately matched with the corresponding sound source and its propagation path, that is, according to the starting and ending relationship of the sound line propagation path, each diffraction attenuation coefficient is spatially associated with the corresponding sound source position and receiving area to form structured sound source-attenuation association data to ensure that the energy attenuation on each propagation path can be accurately attributed to the interaction process between its corresponding sound source and obstacle. Next, a sound pressure level superposition calculation is performed based on the source-attenuation correlation data. This involves logarithmically superimposing the sound pressure levels from different sound sources and different propagation paths (including direct, reflected, and diffracted paths) at the same spatial grid point according to the principle of energy addition. This involves first converting the decibel values ​​to sound intensity, summing them, and then converting them back to decibels. This yields a preliminary superimposed sound pressure level for that point, which preliminarily reflects the combined effect of multi-source noise in space. However, this preliminary result does not yet account for the influence of complex environmental factors such as ground absorption, atmospheric attenuation, wind speed gradient, and vegetation absorption. Therefore, environmental correction is required for the preliminary superimposed sound pressure level. Using the surface material, greenery distribution, and meteorological parameters contained in the geographic information data, a corresponding correction model (such as the ground absorption correction term in ISO 9613-2) is introduced to calculate the additional attenuation at each grid point. This is then deducted from the preliminary superimposed sound pressure level to obtain a more realistic corrected superimposed sound pressure level.Subsequently, considering the significant time-varying characteristics of environmental noise, such as peak and flat fluctuations in traffic flow and the intermittent nature of construction activities, the time characteristics of the corrected superimposed sound pressure level need to be integrated. That is, combining the time series characteristics of the original noise monitoring data, the sound pressure level results are time-weighted (such as using the time statistical weight of the equivalent continuous A sound level Leq or the maximum sound level Lmax) to generate time-weighted superimposed data reflecting the long-term trend or instantaneous peak of noise. Finally, to facilitate subsequent map updates and visualization, the time-weighted superimposed data is spatially formatted. That is, it is mapped to a unified geographic grid coordinate system and output using a standard data format (such as GeoTIFF or gridded CSV) to form a standardized dataset with spatial coordinates, sound pressure level values, and time attributes, namely, noise distribution superimposed data. For example, in the application scenario of urban road intersections, when multiple buses stop and start intensively during the morning rush hour, the system converts the radiation power of each vehicle into sound pressure level through the above process, matches it with the attenuation coefficient caused by diffraction at the edge of the building, and corrects the influence of ground reflection and green belt sound absorption after superposition. It is weighted according to the one-hour equivalent sound level and finally generates high-resolution noise distribution superposition data reflecting the noise hotspot areas during that period, providing an accurate, dynamic and uniformly formatted input basis for updating the original noise map.

[0040] The above describes the noise map updating method based on noise monitoring data in the embodiment of the present invention. The following describes the noise map updating device based on noise monitoring data in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a noise map updating device based on noise monitoring data includes: The acquisition module 21 is used to collect environmental noise in the target area through an acoustic sensor to obtain raw noise data, and perform time-frequency domain decomposition on the raw noise data to obtain a noise time-frequency feature vector; An analysis module 22 is configured to perform sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; a calculation module 23 for calculating a sound ray propagation path of the ambient noise in a target area based on the noise source spatial distribution data, and performing a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; A superposition module 24 is configured to perform sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; The updating module 25 is configured to incrementally update the preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

[0041] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0042] Reference Figure 3 The embodiment of the present invention provides a device for generating a virtual human lecture video, comprising: at least one processor 301; At least one memory 302, configured to store at least one program; When at least one program is executed by at least one processor 301, the at least one processor 301 implements a method for generating a virtual human lecture video.

[0043] Similarly, the contents of the above method embodiments are applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0044] In some optional embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided in an exemplary manner for the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operation and logic flow presented herein. Optional embodiments are contemplated in which the order of the various operations is changed and the sub-operations described as a part of a larger operation are performed independently.

[0045] Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise indicated, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It will also be understood that a detailed discussion of the actual implementation of each module is not necessary for understanding the present invention. More specifically, given the properties, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the module will be understood within the ordinary skill of an engineer. Therefore, a person skilled in the art using ordinary skill will be able to implement the present invention set forth in the claims without undue experimentation. It will also be understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0046] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several programs for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0047] The logic and / or steps represented in a flowchart or otherwise described herein, for example, may be considered as an ordered list of executable programs for implementing the logical functions, and may be embodied in any computer-readable medium for use by, or in conjunction with, a program execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can retrieve and execute a program from a program execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" may be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, a program execution system, apparatus, or device.

[0048] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0049] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable program execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0050] In the above description of this specification, reference to the terms "one embodiment / example," "another embodiment / example," or "certain embodiments / examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0051] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

[0052] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A noise map updating method based on noise monitoring data, characterized in that: The following steps are involved: The acoustic sensor is used to collect ambient noise in the target area to obtain raw noise data, and the raw noise data is decomposed in the time-frequency domain to obtain a noise time-frequency feature vector; Performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; Calculating a sound ray propagation path of the ambient noise in a target area based on the noise source spatial distribution data, and performing a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; Performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; An incremental update is performed on the preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

2. The noise map updating method based on noise monitoring data according to claim 1, characterized in that: The step of performing time-frequency domain decomposition on the original noise data to obtain a noise time-frequency feature vector includes: Performing short-time Fourier transform on the original noise data to obtain time-spectrogram data, and counting frequency band energy in the time-spectrogram data; Performing multi-scale analysis on the frequency band energy by wavelet packet decomposition to obtain the noise fluctuation coefficient; The noise fluctuation coefficient is subjected to feature dimensionality reduction processing to obtain a noise time-frequency feature vector.

3. The noise map updating method based on noise monitoring data according to claim 1, characterized in that: The performing sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data includes: Extracting the noise characteristic peak value of each channel in the noise time-frequency feature vector, and calculating the time difference of arrival of sound sources between different channels based on the noise characteristic peak value; Constructing a positioning equation group based on the time difference of arrival of the sound source and the spatial coordinates of the acoustic sensor, and iteratively solving the positioning equation group to obtain the coordinates of the sound source; Performing density cluster analysis on the sound source coordinates to obtain sound source clusters, and performing center coordinate calculation on the sound source clusters to obtain cluster center coordinates; The radiation power and influence range of each sound source are calculated based on the cluster center coordinates, and the cluster center coordinates are associated and integrated with the radiation power and influence range to obtain noise source spatial distribution data.

4. The noise map updating method based on noise monitoring data according to claim 1, characterized in that: The calculating the sound ray propagation path of the ambient noise in the target area based on the noise source spatial distribution data includes: Determining a sound ray emission source based on cluster center coordinates and radiation power in the noise source spatial distribution data, and discretizing the emission direction angle of the sound ray emission source to obtain a sound ray direction parameter; Calling geographic information data of the target area, and performing spatial intersection detection on the geographic information data and the sound line direction parameter to obtain a sound line-geographic action point; The reflection and refraction parameters of the ambient noise in the target area are calculated for the sound line-geographical action point, and the path of the sound line-geographical action point is extended based on the reflection and refraction parameters to obtain the sound line propagation path.

5. The noise map updating method based on noise monitoring data according to claim 1, characterized in that: The performing diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient includes: Performing spatial tangency detection on the sound ray propagation path and the edge of the obstacle in the target area to obtain a diffraction critical point set, and performing Fresnel zone occlusion analysis on the diffraction critical point set to obtain a Fresnel zone parameter value; Calculating a diffraction coefficient base value based on the Fresnel zone parameter value, and performing energy attenuation conversion on the diffraction coefficient base value and the diffraction critical point set to obtain a diffraction energy loss value; The diffraction effect of the ambient noise in the target area is calculated based on the diffraction energy loss value to obtain a diffraction attenuation coefficient.

6. The noise map updating method based on noise monitoring data according to claim 1, characterized in that: The performing sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data includes: Performing sound pressure level conversion on the radiation power in the noise source spatial distribution data to obtain sound source sound pressure level data, and performing spatial correlation matching on the sound source sound pressure level data and the diffraction attenuation coefficient to obtain sound source-attenuation correlation data; Performing a sound pressure level superposition calculation based on the sound source-attenuation correlation data to obtain a preliminary superposition sound pressure level, and performing environmental correction on the preliminary superposition sound pressure level to obtain a corrected superposition sound pressure level; The time characteristics of the corrected superimposed sound pressure levels are integrated to obtain time-weighted superimposed data, and the time-weighted superimposed data are formatted for spatial distribution to obtain noise distribution superimposed data.

7. A noise map updating device based on noise monitoring data, characterized in that: include: An acquisition module is used to collect ambient noise in a target area through an acoustic sensor to obtain raw noise data, and to perform time-frequency domain decomposition on the raw noise data to obtain a noise time-frequency feature vector; An analysis module is used to perform sound source localization analysis on the target area based on the noise time-frequency feature vector to obtain noise source spatial distribution data; a calculation module, configured to calculate a sound ray propagation path of the ambient noise in a target area based on the noise source spatial distribution data, and perform a diffraction effect analysis on obstacles in the target area based on the sound ray propagation path to obtain a diffraction attenuation coefficient; a superposition module, configured to perform sound field superposition on the diffraction attenuation coefficient and the noise source spatial distribution data to obtain noise distribution superposition data; An updating module is used to incrementally update a preset original noise map based on the noise distribution superposition data to obtain an updated regional noise map.

8. A noise map updating device based on noise monitoring data, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the noise map updating method based on noise monitoring data according to any one of claims 1 to 6.

9. A storage medium storing a program executable by a processor, characterized in that: The processor-executable program is used to implement a noise map updating method based on noise monitoring data as claimed in any one of claims 1 to 6 when executed by the processor.

Images