Multi-source data fusion-based quiet cell noise map construction and evaluation method

Abstract

The invention relates to a multi-source data fusion-based Ningxia cell noise map construction and evaluation method, which comprises the following steps of: establishing a frequency-variable sound absorption parameter matrix of a building surface based on multi-spectral scanning data and an acoustic database, and combining a burst noise voiceprint recognition result and traffic monitoring data, generating a burst noise influence map through the burst noise space-time correlation model; the real-time sound pressure data and the frequency-variable sound absorption parameter matrix are fused, and a sound field is dynamically reconstructed to generate a noise thermodynamic diagram; and based on a Zwicker loudness model, coupling a burst noise influence map and resident electrocardiogram monitoring data, calculating a composite noise dysphoria index, constructing a noise map in combination with a noise thermodynamic diagram, and carrying out comprehensive evaluation by adopting an entropy weight-TOPSIS algorithm to obtain a noise evaluation result. According to the method, the noise map is constructed through multi-source heterogeneous data fusion, dynamic sound field modeling and physiological and psychological coupling calculation, and evaluation is performed in combination with multi-dimensional indexes, so that precise evaluation of a cell noise environment and scientific dynamic optimization of a noise reduction scheme are realized, and objectivity and pertinence of noise treatment are effectively improved.

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and particularly relates to a method for constructing and evaluating a quiet community noise map based on multi-source data fusion. Background Art

[0002] With the development of architectural acoustics and noise control technologies, a variety of acoustic modeling and noise analysis technologies have emerged, which can accurately simulate and predict the noise propagation, distribution, and impact on humans in complex environments, and thus can provide a scientific basis for noise control and environmental planning. However, traditional residential area noise assessment methods mainly rely on the equivalent sound level measurement at fixed monitoring points and static noise map modeling. Although the noise distribution can be estimated through interpolation of historical sound pressure data or a simplified geometric acoustics model, and the noise reduction area can be delimited in combination with expert experience, the equivalent sound level measurement at fixed monitoring points cannot reflect the dynamic changes of noise in real time, it is difficult to capture the impact of sudden noise in residential areas, and there is a lack of comprehensive assessment of the psycho-physiological impact of noise, making it difficult to comprehensively reflect the actual impact of noise on residents. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method for constructing and evaluating a quiet community noise map based on multi-source data fusion to further improve the accuracy of residential area noise assessment.

[0004] In a first aspect, the present application provides a method for constructing and evaluating a quiet community noise map based on multi-source data fusion, including:

[0005] Performing acoustic heterogeneity modeling processing on the building surface according to multi-spectral scanning data and a preset acoustic database to obtain a frequency-variable sound absorption parameter matrix;

[0006] Calculating through a sudden noise spatio-temporal correlation model according to the voiceprint recognition result of sudden noise and traffic monitoring data to obtain a sudden noise impact map;

[0007] Performing dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with real-time sound pressure data to generate a noise heat map;

[0008] Based on the Zwicker loudness model, performing psycho-physiological coupling calculation processing according to the sudden noise impact map combined with electrocardiogram monitoring data, and outputting a composite noise annoyance index;

[0009] Constructing a noise map according to the composite noise annoyance index and the noise heat map, and using an entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision-making scheme.

[0010] In one embodiment, based on the multispectral scanning data and a preset acoustic database, acoustic heterogeneity modeling processing is performed on the building surface to obtain a frequency-variable sound absorption parameter matrix, including:

[0011] Identify the building facade material types according to the multispectral scanning data through a spectral matching algorithm, and generate a material distribution matrix;

[0012] Based on the preset acoustic database, match the frequency-variable sound absorption coefficients of each material type, and combine with the material distribution matrix to construct a three-dimensional sound absorption parameter field;

[0013] Through the three-dimensional sound absorption parameter field, use the ray tracing method to calculate the noise incident angle, correct the frequency-variable sound absorption coefficient, and obtain the frequency-variable sound absorption parameter matrix.

[0014] In one embodiment, according to the voiceprint recognition result of the sudden noise and the traffic monitoring data, calculate through a sudden noise spatio-temporal correlation model to obtain a sudden noise impact map, including:

[0015] According to the voiceprint spectrum characteristics of the voiceprint recognition result of the sudden noise, perform transient event classification processing and output the sound source type label;

[0016] According to the YOLOv5 model and the traffic monitoring data, perform vehicle trajectory tracking processing to generate a noise source spatio-temporal coordinate sequence;

[0017] Input the noise source spatio-temporal coordinate sequence and the sound source type label into the sudden noise spatio-temporal correlation model for calculation to obtain a comprehensive influence value;

[0018] Map the comprehensive influence value to the grid of the geographic information system to generate a sudden noise impact map.

[0019] In one embodiment, calculate the comprehensive influence value through the following formula:

[0020]

[0021] Among them, ΔL(t,x,y) is the comprehensive influence value of the sudden noise, A k is the weight of the sound source type label, β is the terrain attenuation coefficient, τ is the time attenuation factor, SPL k is the sound pressure level of the sound source of the kth sudden noise, (x k (t),y k (t)) is the noise source spatio-temporal coordinate sequence at time t, and t is the occurrence time of the sound source of the sudden noise.

[0022] In one embodiment, based on the Zwicker loudness model, perform psycho-physiological coupling calculation processing according to the sudden noise impact map combined with electrocardiogram monitoring data, and output a composite noise annoyance index, including:

[0023] Calculate the equivalent sound energy through time integration according to the sudden noise impact spectrum;

[0024] Calculate the instantaneous loudness based on the equivalent sound energy through the Zwicker loudness model;

[0025] Perform heart rate variability power spectrum analysis on the electrocardiogram monitoring data and calculate the HRV energy ratio;

[0026] Calculate through convolution integral based on the HRV energy ratio and the instantaneous loudness to obtain the composite noise annoyance index.

[0027] In one embodiment, perform dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with the real-time sound pressure data to generate a noise heat map, including:

[0028] Input the frequency-variable sound absorption parameter matrix into the pre-trained finite element sound field basis function to extract sound field features;

[0029] Based on the sound field features, use the meta-learning network to dynamically allocate the weights of the basis function to generate the initial sound pressure distribution output by the reduced-order model;

[0030] Calibrate the reduced-order model through the real-time sound pressure data and the initial sound pressure distribution, update the weights, and obtain the updated sound pressure distribution;

[0031] Map the updated sound pressure distribution to the geographic information system grid through Kriging spatial interpolation to generate a noise heat map.

[0032] In one embodiment, construct a noise map according to the composite noise annoyance index and the noise heat map, and use the entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision scheme, including:

[0033] Construct a noise map according to the composite noise annoyance index and the noise heat map. The geographic grid points in the noise map include impact index data, and the impact index data includes physiological impact indicators, psychological perception indicators, and functional adaptability indicators;

[0034] Based on the entropy weight-TOPSIS fusion algorithm, perform standardization processing on the impact index data in the noise map, calculate the information entropy of each impact index, and allocate the weights of each impact index according to the information entropy;

[0035] Determine the ideal values of each impact index, and calculate the distance between the corresponding area and the ideal value based on the weights to obtain the noise evaluation result.

[0036] In a second aspect, the present application also provides a quiet community noise map construction and evaluation system based on multi-source data fusion. The system includes:

[0037] A parameter generation module, configured to:

[0038] Perform acoustic heterogeneity modeling processing on the building surface according to the multispectral scanning data and the preset acoustic database to obtain a frequency-variable sound absorption parameter matrix;

[0039] Calculate through the sudden noise spatio-temporal correlation model according to the voiceprint recognition result of the sudden noise and the traffic monitoring data to obtain a sudden noise impact map;

[0040] A sound field reconstruction module, configured to perform dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with the real-time sound pressure data to generate a noise heat map;

[0041] An annoyance quantification module, configured to perform psychophysiological coupling calculation processing based on the Zwicker loudness model according to the sudden noise impact map combined with the electrocardiogram monitoring data, and output a composite noise annoyance index;

[0042] A noise evaluation module, configured to construct a noise map according to the composite noise annoyance index and the noise heat map, and adopt an entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision scheme.

[0043] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps in the first aspect are implemented.

[0044] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the first aspect are implemented.

[0045] The above method for constructing and evaluating a quiet community noise map based on multi-source data fusion performs acoustic heterogeneity modeling of building surfaces by using multi-spectral scanning data and a preset acoustic database, overcomes the deficiencies of traditional static modeling methods, can update the acoustic characteristics of building surfaces in real time, and thus provides more accurate data support for subsequent sound field reconstruction. By using the voiceprint recognition results of sudden noises and traffic monitoring data to input into a spatio-temporal correlation model to calculate the influence range and propagation path of sudden noises, it can not only quickly locate the noise source but also track the dynamic changes of the noise in real time, avoiding the limitation that fixed monitoring points cannot capture sudden noises. Secondly, by combining a frequency-variable sound absorption parameter matrix and real-time sound pressure data to perform dynamic sound field reconstruction processing, this method can reflect the distribution of noise in the community in real time and generate a noise heat map. In addition, this method also performs psycho-physiological coupling calculations through the Zwicker loudness model, combined with the sudden noise influence map and electrocardiogram monitoring data, to more comprehensively evaluate the actual impact of noise on residents and obtain a composite noise annoyance index. Finally, this method constructs a noise map based on the composite noise annoyance index and the noise heat map, and uses an entropy weight-TOPSIS fusion algorithm for multi-dimensional index comprehensive evaluation. It not only considers the physical distribution of noise but also combines psycho-physiological effects, and can more scientifically evaluate the community noise environment. Compared with traditional community noise assessment methods, this method greatly improves the accuracy and scientific nature of noise assessment through dynamic modeling, psycho-physiological coupling calculations, and multi-dimensional comprehensive evaluation, providing an effective scientific basis for community noise control. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0047] Figure 1 Flowchart of the method for constructing and evaluating a quiet community noise map based on multi-source data fusion provided by an exemplary embodiment of the present invention;

[0048] Figure 2 Schematic structural diagram of the system for constructing and evaluating a quiet community noise map based on multi-source data fusion provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the following further details the present application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] In one embodiment, as Figure 1 shown, a method for constructing and evaluating a quiet community noise map based on multi-source data fusion is provided. In this embodiment, an example is given where this method is applied to a terminal. It can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0051] S101: Perform acoustic heterogeneity modeling processing on the building surface according to the multispectral scanning data and the preset acoustic database to obtain a frequency-variable sound absorption parameter matrix.

[0052] Specifically, the building surface in the community can be scanned by using a multispectral scanning device to obtain data on the reflection, absorption, and other characteristics of the building surface at different wavelengths. This multispectral scanning data can reflect the microscopic structure and physical properties of the building surface material. And the preset acoustic database stores the sound absorption performance parameters of different materials at different frequencies. By matching and fitting the multispectral scanning data with the parameters in the preset acoustic database, a frequency-variable sound absorption parameter matrix of the community building surface can be obtained. This frequency-variable sound absorption parameter matrix provides accurate boundary conditions for subsequent sound field reconstruction and can further improve the accuracy of noise distribution simulation.

[0053] S102: Calculate according to the voiceprint recognition result of the sudden noise and the traffic monitoring data through the sudden noise spatio-temporal correlation model to obtain a sudden noise impact map.

[0054] Specifically, a pre-trained neural network model can be constructed to extract features of sudden noises such as vehicle horns and construction machinery, and then the voiceprint information of the noise source, that is, the voiceprint recognition result, can be obtained. And through the traffic monitoring data, spatio-temporal information such as the location, moving direction, and speed of the noise source can be obtained. By using the sudden noise spatio-temporal correlation model, the voiceprint recognition result can be matched with the timestamp and location information in the traffic monitoring data, and then the propagation path and influence range of the sudden noise can be traced, and a sudden noise impact map can be generated. This map can intuitively display the intensity change and propagation path of the sudden noise at different positions, thus avoiding the limitation that fixed monitoring points cannot capture sudden noises.

[0055] S103: Perform dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with the real-time sound pressure data to generate a noise heat map.

[0056] Specifically, real-time sound pressure data can be collected in real time through a sound pressure sensor network deployed in the community. This real-time sound pressure data reflects the noise intensity at various locations in the community at the current moment. Based on the frequency-variable sound absorption parameter matrix and the real-time sound pressure data, dynamic sound field reconstruction algorithms such as the finite element method, the boundary element method, or the ray tracing method can be used to simulate the propagation process of sound waves in a complex building environment, calculate the reflection, refraction, and absorption effects of sound waves, and thus reconstruct the sound field distribution in the community at the current moment, and then generate a noise heat map with high spatio-temporal resolution. This noise heat map can visually display the noise intensity distribution in each area of the community in the form of colors or numerical values, and reflect the dynamic changes of noise in real time.

[0057] S104: Based on the Zwicker loudness model, perform psycho-physiological coupling calculation and processing according to the sudden noise impact map combined with electrocardiogram monitoring data, and output a composite noise annoyance index.

[0058] Specifically, the Zwicker loudness model can combine the physical characteristics of noise with the subjective perception of the human ear, while the electrocardiogram monitoring data reflects the impact of noise on the physiological state of the human body. Through psycho-physiological coupling calculation, a composite noise annoyance index is output. Based on the Zwicker loudness model, combined with the sudden noise impact map generated in S102 and the real-time collected electrocardiogram monitoring data, psycho-physiological coupling calculation can more comprehensively evaluate the actual impact of noise on residents and obtain a composite noise annoyance index. This composite noise annoyance index comprehensively considers the physical characteristics of noise and the physiological reactions of the human body, can more accurately reflect the actual impact degree of noise on people, and provides a more scientific and objective index for subsequent noise evaluation.

[0059] S105: Construct a noise map based on the composite noise annoyance index and the noise heat map, and use the entropy weight-TOPSIS fusion algorithm to perform comprehensive evaluation processing of multi-dimensional indicators through the noise map to obtain a noise evaluation result, which is used to generate a noise reduction priority decision plan.

[0060] Specifically, a noise map can be constructed according to the composite noise annoyance index obtained in S104 and the noise heat map generated in S103. This noise map combines the physical distribution of noise with the degree of impact on the human body, and can comprehensively display the current situation of noise pollution. And by using the entropy weight-TOPSIS fusion algorithm, the weights of each indicator in the noise map can be dynamically determined by the entropy weight method, that is, comprehensively considering multiple impact indicators such as noise intensity, distribution range, and impact on the human body, and using the TOPSIS method for comprehensive evaluation to obtain a noise evaluation result. This noise evaluation result provides a scientific basis for community noise control and helps to formulate targeted and prioritized noise reduction strategies.

[0061] In one embodiment, based on the multispectral scanning data and a preset acoustic database, building surface acoustic heterogeneity modeling processing is performed to obtain a frequency-variable sound absorption parameter matrix, including:

[0062] Identify the building facade material types according to the multispectral scanning data through a spectral matching algorithm, and generate a material distribution matrix;

[0063] Based on the preset acoustic database, match the frequency-variable sound absorption coefficients of each material type, and combine with the material distribution matrix to construct a three-dimensional sound absorption parameter field;

[0064] Through the three-dimensional sound absorption parameter field, use the ray tracing method to calculate the noise incident angle, correct the frequency-variable sound absorption coefficient, and obtain the frequency-variable sound absorption parameter matrix.

[0065] Specifically, through a spectral matching algorithm such as support vector machine classification, the scanned spectral data can be compared with the preset material spectral database. By analyzing the reflection characteristics at different wavelengths, the material types of the building facade can be identified. And according to the identified material types, a material distribution matrix is generated. Then, according to the material types recorded in the material distribution matrix, the corresponding frequency-variable sound absorption coefficients can be matched from the preset acoustic database, and combined with the material distribution matrix to construct a three-dimensional sound absorption parameter field. Among them, the three-dimensional sound absorption parameter field is a spatially distributed parameter field, which can describe the sound absorption performance of the building surface at different positions and different frequencies, and can provide accurate boundary conditions for subsequent sound field simulation. Since the sound absorption characteristics of materials change with the incident angle, through the constructed three-dimensional sound absorption parameter field, the ray tracing method can be used to simulate the propagation path of sound waves in the building environment and calculate the incident angles of sound waves at different positions. Then, according to the calculated incident angles, the frequency-variable sound absorption coefficient is corrected. The corrected frequency-variable sound absorption coefficient can more accurately describe the actual sound absorption performance of noise on the building surface. Finally, the corrected frequency-variable sound absorption coefficients can be organized into a frequency-variable sound absorption parameter matrix. This matrix comprehensively considers the influence of material types, spatial distribution, and sound wave incident angles, provides accurate boundary conditions for subsequent sound field reconstruction, and can significantly improve the accuracy of noise simulation. Schematically, the ray tracing method is a method that regards the sound wave propagation process as ray propagation, and can obtain the sound field distribution by tracing the propagation path of rays in space.

[0066] In one embodiment, according to the voiceprint recognition result of the sudden noise and the traffic monitoring data, calculations are performed through a sudden noise spatio-temporal correlation model to obtain a sudden noise impact map, including:

[0067] Perform transient event classification processing according to the voiceprint spectral characteristics of the voiceprint recognition result of the sudden noise, and output a sound source type label;

[0068] Based on the YOLOv5 model and traffic monitoring data, vehicle trajectory tracking is performed to generate a spatio-temporal coordinate sequence of noise sources;

[0069] The spatio-temporal coordinate sequence of noise sources and the sound source type label are input into the spatio-temporal correlation model of impulsive noise for calculation to obtain a comprehensive influence value;

[0070] The comprehensive influence value is mapped into the grid of the geographic information system to generate an impulsive noise influence map.

[0071] Specifically, a pre-trained deep learning model such as a convolutional neural network can be used to extract features of impulsive noise to obtain the soundprint spectrum features. The soundprint spectrum features include the spectrum information of the soundprint and can reflect the unique features of the noise. Secondly, through transient event classification processing, the extracted soundprint spectrum features are compared with known sound source types to identify the source type of the noise, such as vehicle honking, construction machinery, etc., and the sound source type label is output, providing a basis for subsequent noise source localization and impact assessment. Among them, YOLOv5 is an efficient real-time object detection algorithm that can quickly and accurately identify and locate targets such as vehicles and pedestrians in videos. The traffic monitoring data can be analyzed in real time through the YOLOv5 model, and through vehicle trajectory tracking processing, the position information of the vehicle at different time points is recorded to generate a spatio-temporal coordinate sequence of noise sources. The spatio-temporal coordinate sequence of noise sources can then describe in detail the distribution and movement trajectory of the noise source in time and space. The spatio-temporal correlation model of impulsive noise is a mathematical model that combines time and space information and can analyze the propagation characteristics of noise sources at different positions and times. The spatio-temporal coordinate sequence of noise sources and the sound source type label can be input into this model, and the comprehensive influence value is obtained through calculation. The comprehensive influence value comprehensively considers the physical characteristics of the noise and the propagation environment and can more accurately evaluate the actual impact of the noise.

[0072] In one embodiment, the comprehensive influence value is calculated by the following formula:

[0073]

[0074] where, ΔL(t,x,y) is the comprehensive influence value of the impulsive noise, A k is the weight of the sound source type label, β is the terrain attenuation coefficient, τ is the time attenuation factor, SPL k is the sound pressure level of the sound source of the k-th impulsive noise, (x k (t),y k (t)) is the spatio-temporal coordinate sequence of the noise source at time t, and t is the occurrence time of the sound source of the impulsive noise.

[0075] Specifically, the influence weights of different types of sudden noises can be initialized through a preset acoustic database and a noise type weight table. Schematically, the weight of vehicle horn honking can be set as A1 = 1.2, while the weight of construction machinery impact is A2 = 1.5, to reflect the differences in the influence of different types of noises on human perception. In addition, terrain features such as building height, density, and green belt distribution can be extracted based on the three-dimensional geographic information system data of the community, and the terrain attenuation coefficient β can be calibrated through corresponding acoustic propagation simulation experiments. For example, β = 0.05 can indicate that the attenuation of noise in an open area with the propagation distance is small, and β = 0.15 can indicate that the attenuation of noise in a densely built area with the propagation distance is large. The time attenuation factor τ is obtained by fitting historical noise event data. For example, τ = 0.1 can indicate that the influence of noise decays exponentially in the time dimension. Through It can indicate that the influence of noise decays exponentially over time.

[0076] In one embodiment, based on the Zwicker loudness model, psycho-physiological coupling calculation and processing are performed according to the sudden noise influence map combined with electrocardiogram monitoring data, and a composite noise annoyance index is output, including:

[0077] The equivalent sound energy is calculated by time integration according to the sudden noise influence map;

[0078] The instantaneous loudness is calculated based on the equivalent sound energy through the Zwicker loudness model;

[0079] The HRV energy ratio is calculated by performing heart rate variability power spectrum analysis according to the electrocardiogram monitoring data;

[0080] The composite noise annoyance index is obtained by calculating through convolution integral according to the HRV energy ratio and the instantaneous loudness.

[0081] Specifically, equivalent sound energy refers to the total energy of noise within a certain period of time. The intensity distribution data of noise on its time series, such as the grid point noise value ΔL(t,x,y)), can be extracted through the sudden noise impact spectrum, and the equivalent sound energy at each position can be calculated using the time-energy integration algorithm. Thus, the dynamic noise impact can be quantified as the cumulative sound energy, providing input for subsequent loudness calculation. Moreover, the random fluctuations of instantaneous noise can be eliminated through time integration, and the cumulative effect of noise energy can be extracted, which is more in line with the long-term perception characteristics of the human body towards noise. The Zwicker loudness model is a model based on the auditory characteristics of the human ear, which can convert the sound pressure level into the loudness level and is more in line with the actual physiological response compared with the traditional A-weighted sound level. In addition, the power spectrum analysis of heart rate variability can evaluate the activity of the autonomic nervous system by analyzing the electrocardiogram monitoring data, that is, the changes in the RR interval in the electrocardiogram data. Schematically, the power spectrum analysis method can be used to decompose the HRV (Heart Rate Variability) signal into different frequency components and calculate the energy ratio of each frequency component, that is, the HRV energy ratio. And the composite noise annoyance index can be obtained by performing convolution integration on the HRV energy ratio and the instantaneous loudness. This index not only reflects the intensity and propagation characteristics of the noise, but also takes into account the psychological and physiological effects of the noise on people, and can more accurately reflect the actual impact degree of the noise on people, providing a more scientific and objective index for noise evaluation.

[0082] In one embodiment, the frequency-variable sound absorption parameter matrix is combined with the real-time sound pressure data to perform dynamic sound field reconstruction processing to generate a noise heat map, including:

[0083] Input the frequency-variable sound absorption parameter matrix into the pre-trained finite element sound field basis function to extract the sound field features;

[0084] Based on the sound field features, use the meta-learning network to dynamically allocate the weights of the basis function to generate the initial sound pressure distribution output by the reduced-order model;

[0085] Calibrate the reduced-order model through the real-time sound pressure data and the initial sound pressure distribution, update the weights, and obtain the updated sound pressure distribution;

[0086] Map the updated sound pressure distribution to the geographical information system grid through Kriging spatial interpolation to generate a noise heat map.

[0087] Specifically, the finite element sound field basis function is a numerical analysis method that can discretize complex sound field problems into a series of simple element problems for solution. By inputting the frequency-varying sound absorption parameter matrix into the pre-trained finite element sound field basis function, sound field characteristics such as sound pressure distribution and sound intensity distribution can be extracted. The meta-learning network is a learning model that can quickly adapt to new tasks. By continuously learning during the training process, it can quickly adjust the model parameters to adapt to new inputs. Through the meta-learning network, the attention weights of each basis function can be calculated based on the correlation between the sound field characteristics and the basis function, and the basis functions can be superimposed according to the weights to generate the initial sound pressure distribution of the reduced-order model. Among them, the reduced-order model is a simplified model that can greatly reduce the computational amount and improve the computational efficiency while ensuring a certain accuracy. By comparing the real-time sound pressure data with the initial sound pressure distribution, the weights of the reduced-order model can be updated through calibration algorithms such as the least squares method. The updated weights can more accurately reflect the distribution of the actual sound field, thereby obtaining a more accurate sound pressure distribution. In addition, Kriging spatial interpolation is a geostatistical method that can predict the values of unknown points based on the measured values of known points. Through Kriging interpolation, the updated sound pressure distribution data can be converted into color information in a geographic information system grid to generate a noise heat map. The noise heat map can represent the level of noise intensity by the depth of color, and can intuitively display the distribution of noise in space.

[0088] In one embodiment, a noise map is constructed based on the composite noise annoyance index and the noise heat map, and the entropy weight-TOPSIS fusion algorithm is used to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision-making scheme, including:

[0089] Based on the composite noise annoyance index and the noise heat map, a noise map is constructed, and the geographical grid points in the noise map include impact index data, and the impact index data includes physiological impact indexes, psychological perception indexes and functional adaptability indexes;

[0090] Based on the entropy weight-TOPSIS fusion algorithm, the impact index data in the noise map is standardized, the information entropy of each impact index is calculated, and the weights of each impact index are allocated according to the information entropy;

[0091] The ideal values of each impact index are determined, and based on the weights, the distances between the corresponding regions and the ideal values are calculated to obtain a noise evaluation result.

[0092] Specifically, the composite noise annoyance index and the noise heat map can be spatially aligned and data - fused to construct a noise map. Among them, the physiological impact index can quantify the impact degree of noise on the autonomic nervous system through the HRV energy ratio. For example, an HRV energy ratio greater than 3.0 is used to represent high physiological stimulation. The psychological perception index can be measured by the composite noise annoyance index, while the functional adaptability index can measure the noise limit value according to regional planning attributes such as residential areas, commercial areas, and schools, and calculate the noise exceeding rate. The three types of indicators are normalized by range to eliminate the dimension difference and ensure the comparability between different indicators. And the entropy method is used to calculate the information entropy of each impact indicator. The larger the information entropy, the richer the information of the indicator and the greater the weight. The maximum or minimum value of each indicator can be taken to determine the ideal value of each impact indicator. Subsequently, the TOPSIS method can be used to calculate the distance between each region and the ideal value, and then calculate the relative proximity of each region to the ideal value to obtain the noise evaluation result, thereby enabling a dual evaluation of the noise environment on human perception and regional function. Schematically, this distance can be the Euclidean distance.

[0093] Based on the same inventive concept, as Figure 2 shown, the present application also provides a quiet community noise map construction and evaluation system 200 based on multi - source data fusion. The system includes:

[0094] A parameter generation module 201, configured to:

[0095] Perform building surface acoustic heterogeneity modeling processing based on multi - spectral scanning data and a preset acoustic database to obtain a frequency - variable sound absorption parameter matrix;

[0096] Perform calculations through a sudden noise spatio - temporal correlation model based on the voiceprint recognition result of sudden noise and traffic monitoring data to obtain a sudden noise impact map;

[0097] A sound field reconstruction module 202, configured to perform dynamic sound field reconstruction processing based on the frequency - variable sound absorption parameter matrix combined with real - time sound pressure data to generate a noise heat map;

[0098] An annoyance quantification module 203, configured to perform psycho - physiological coupling calculation processing based on the Zwicker loudness model, combine the sudden noise impact map with electrocardiogram monitoring data, and output a composite noise annoyance index;

[0099] A noise evaluation module 204, configured to construct a noise map based on the composite noise annoyance index and the noise heat map, and adopt an entropy weight - TOPSIS fusion algorithm to perform multi - dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision scheme.

[0100] In the above-mentioned quiet community noise map construction and evaluation system based on multi-source data fusion, specifically, the parameter generation module 201 performs building surface acoustic heterogeneity modeling on multi-spectral scanning data and a preset acoustic database to obtain a frequency-variable sound absorption parameter matrix. This matrix can accurately reflect the sound absorption characteristics of the building surface at different frequencies and provides a key acoustic parameter basis for subsequent sound field reconstruction. And based on the voiceprint recognition result of the sudden noise and traffic monitoring data, the sudden noise impact map can be calculated through the sudden noise spatio-temporal correlation model, so as to display the distribution and influence range of the sudden noise in time and space, providing an important basis for the subsequent quantitative analysis of the noise impact. The sound field reconstruction module 202 can then perform dynamic sound field reconstruction processing based on the frequency-variable sound absorption parameter matrix combined with real-time sound pressure data to generate a noise heat map. This process can intuitively present the distribution of noise in the community, including the intensity and spatial position of the noise, helping to understand the overall situation of the noise in the community and providing a visual basis for the evaluation and management of the noise. And the annoyance quantification module 203 can perform psychophysiological coupling calculation processing based on the Zwicker loudness model, combining the sudden noise impact map and electrocardiogram monitoring data, and output a composite noise annoyance index. This index comprehensively considers the physical characteristics of the noise and the psychophysiological response of the human body to the noise, and can more accurately measure the annoyance degree of residents in the noise environment, providing a more comprehensive and scientific index for the evaluation of the impact of noise on residents. In addition, the noise evaluation module 204 constructs a noise map based on the composite noise annoyance index and the noise heat map, and uses the entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result. This result can not only comprehensively consider multiple aspects of the noise, providing a comprehensive and objective evaluation of the community noise situation, but also generate a corresponding noise reduction priority decision-making scheme, and then can specifically determine the areas and measures that need to be given priority to noise reduction treatment, improving the efficiency and effect of the noise reduction work and effectively improving the acoustic environment quality of the community.

[0101] Further, the parameter generation module 201 includes a community heterogeneous surface modeling sub-unit for:

[0102] Identifying the building facade material type according to the multi-spectral scanning data through a spectral matching algorithm to generate a material distribution matrix;

[0103] Based on the preset acoustic database, matching the frequency-variable sound absorption coefficients of each material type, and combining with the material distribution matrix, constructing a three-dimensional sound absorption parameter field;

[0104] Through the three-dimensional sound absorption parameter field, using the ray tracing method to calculate the noise incident angle, correcting the frequency-variable sound absorption coefficient, and obtaining the frequency-variable sound absorption parameter matrix.

[0105] Further, the parameter generation module 201 also includes a noise spatio-temporal correlation sub-unit for:

[0106] Perform transient event classification processing based on the acoustic spectrum characteristics of the voiceprint recognition result of the burst noise, and output the sound source type label;

[0107] Perform vehicle trajectory tracking processing based on the YOLOv5 model and traffic monitoring data, and generate a spatio-temporal coordinate sequence of the noise source;

[0108] Input the spatio-temporal coordinate sequence of the noise source and the sound source type label into the burst noise spatio-temporal correlation model for calculation to obtain the comprehensive influence value;

[0109] Map the comprehensive influence value to the grid of the geographic information system to generate a burst noise influence map.

[0110] Furthermore, the comprehensive influence value is calculated through the following formula:

[0111]

[0112] Among them, ΔL(t,x,y) is the comprehensive influence value of the burst noise, A k is the weight of the sound source type label, β is the terrain attenuation coefficient, τ is the time attenuation factor, SPL k is the sound pressure level of the sound source of the k-th burst noise, (x k (t),y k (t)) is the spatio-temporal coordinate sequence of the noise source at time t, and t is the occurrence time of the sound source of the burst noise.

[0113] Furthermore, the annoyance quantification module 203 includes:

[0114] A noise quantification subunit, which is used to calculate the equivalent sound energy through time integration according to the burst noise influence map;

[0115] Calculate the instantaneous loudness based on the equivalent sound energy through the Zwicker loudness model;

[0116] A noise coupling psychology subunit, which is used for:

[0117] Perform heart rate variability power spectrum analysis according to the electrocardiogram monitoring data, and calculate the HRV energy ratio;

[0118] Calculate through convolution integral according to the HRV energy ratio and the instantaneous loudness to obtain the composite noise annoyance index.

[0119] Furthermore, the sound field reconstruction module 202 includes:

[0120] A feature extraction subunit, which is used to input the frequency-variable sound absorption parameter matrix into the pre-trained finite element sound field basis function to extract sound field features;

[0121] The sound field adjustment subunit is used to dynamically allocate the weights of basis functions based on the sound field characteristics by using a meta-learning network, and generate the initial sound pressure distribution output by the reduced-order model;

[0122] Calibrate the reduced-order model through real-time sound pressure data and the initial sound pressure distribution, update the weights, and obtain the updated sound pressure distribution;

[0123] Map the updated sound pressure distribution to the geographic information system grid through Kriging spatial interpolation to generate a noise heat map.

[0124] Furthermore, the noise evaluation module 204 includes:

[0125] The map construction subunit is used to construct a noise map according to the composite noise annoyance index and the noise heat map. The geographical grid points in the noise map include impact index data, and the impact index data includes physiological impact indicators, psychological perception indicators, and functional adaptability indicators;

[0126] The noise evaluation subunit is used to standardize the impact index data in the noise map based on the entropy weight-TOPSIS fusion algorithm, calculate the information entropy of each impact index, and allocate the weights of each impact index according to the information entropy;

[0127] Determine the ideal values of each impact index, and calculate the distance between the corresponding area and the ideal value based on the weights to obtain the noise evaluation result.

[0128] In an exemplary embodiment, the present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method for constructing and evaluating a quiet community noise map based on multi-source data fusion of the present application are implemented. A multi-core processor is preferably used to improve the parallel processing ability of the system. Memory: Provide sufficient temporary storage space to support the operation of the program and the processing of data. The memory capacity should be large enough to accommodate a large amount of supply information and computing tasks.

[0129] In an exemplary embodiment, the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for constructing and evaluating a quiet community noise map based on multi-source data fusion of the present application are implemented. The computer-readable storage medium may include: read-only memory (ROM, ReadOnly Memory), random access memory (RAM, Random Access Memory), solid-state drive (SSD, SolidState Drives), or optical disc, etc. Among them, the random access memory may include resistive random access memory (ReRAM, Resistance Random Access Memory) and dynamic random access memory (DRAM, Dynamic RandomAccess Memory).

[0130] The above-described embodiments merely represent several implementation manners of the embodiments of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A method for constructing and evaluating a noise map of a quiet community based on multi-source data fusion, characterized in that, The method includes: Performing acoustic heterogeneity modeling processing on the building surface according to the multispectral scanning data and a preset acoustic database to obtain a frequency-variable sound absorption parameter matrix; Calculating through a sudden noise spatio-temporal correlation model according to the voiceprint recognition result of the sudden noise and traffic monitoring data to obtain a sudden noise impact map; Performing dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with real-time sound pressure data to generate a noise heat map; Based on the Zwicker loudness model, performing psycho-physiological coupling calculation processing according to the sudden noise impact map combined with electrocardiogram monitoring data, and outputting a composite noise annoyance index; Constructing a noise map according to the composite noise annoyance index and the noise heat map, and adopting an entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, and the noise evaluation result is used to generate a noise reduction priority decision plan.

2. The method according to claim 1, characterized in that, The performing acoustic heterogeneity modeling processing on the building surface according to the multispectral scanning data and a preset acoustic database to obtain a frequency-variable sound absorption parameter matrix includes: Identifying the building facade material type through a spectral matching algorithm according to the multispectral scanning data to generate a material distribution matrix; Based on the preset acoustic database, matching the frequency-variable sound absorption coefficients of each material type, and combining with the material distribution matrix to construct a three-dimensional sound absorption parameter field; Through the three-dimensional sound absorption parameter field, using the ray tracing method to calculate the noise incident angle, and correcting the frequency-variable sound absorption coefficient to obtain the frequency-variable sound absorption parameter matrix.

3. The method according to claim 1, wherein The calculating through a sudden noise spatio-temporal correlation model according to the voiceprint recognition result of the sudden noise and traffic monitoring data to obtain a sudden noise impact map includes: Performing transient event classification processing according to the voiceprint spectral characteristics of the voiceprint recognition result of the sudden noise, and outputting a sound source type label; Performing vehicle trajectory tracking processing according to the YOLOv5 model and the traffic monitoring data to generate a sudden noise source spatio-temporal coordinate sequence; Inputting the sudden noise source spatio-temporal coordinate sequence and the sound source type label into the sudden noise spatio-temporal correlation model for calculation to obtain a comprehensive influence value; Mapping the comprehensive influence value to the grid of the geographic information system to generate the sudden noise impact map.

4. The method according to claim 3, wherein The comprehensive influence value is calculated by the following formula: Among them, ΔL(t,x,y) is the comprehensive influence value of the burst noise, A k is the weight of the sound source type label, β is the terrain attenuation coefficient, τ is the time attenuation factor, SPL k is the sound pressure level of the sound source of the k-th burst noise, (x k (t), y k (t)) is the spatio-temporal coordinate sequence of the noise source at time t, and t is the occurrence time of the sound source of the burst noise.

5. The method according to claim 1, wherein The performing psycho-physiological coupling calculation processing based on the Zwicker loudness model according to the sudden noise impact map combined with electrocardiogram monitoring data and outputting a composite noise annoyance index includes: Calculating the equivalent sound energy through time integration according to the sudden noise impact map; Calculating the instantaneous loudness based on the equivalent sound energy through the Zwicker loudness model; Performing heart rate variability power spectrum analysis according to the electrocardiogram monitoring data to calculate the HRV energy ratio; Calculating according to the HRV energy ratio and the instantaneous loudness through convolution integral to obtain the composite noise annoyance index.

6. The method according to claim 1, wherein The performing dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with real-time sound pressure data to generate a noise heat map includes: Inputting the frequency-variable sound absorption parameter matrix into a pre-trained finite element sound field basis function to extract sound field features; Based on the sound field characteristics, a meta-learning network is used to dynamically allocate the weights of the basis functions to generate the initial sound pressure distribution of the reduced-order model output; The reduced-order model is calibrated by the real-time sound pressure data and the initial sound pressure distribution, and the weights are updated to obtain the updated sound pressure distribution; The updated sound pressure distribution is mapped to the geographic information system grid through Kriging spatial interpolation to generate the noise heat map.

7. The method according to claim 1, characterized in that Constructing a noise map according to the composite noise annoyance index and the noise heat map, and using the entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, where the noise evaluation result is used to generate a noise reduction priority decision scheme, including: Construct the noise map according to the composite noise annoyance index and the noise heat map. The geographic grid points in the noise map include impact index data, and the impact index data includes physiological impact indicators, psychological perception indicators, and functional adaptability indicators; Based on the entropy weight-TOPSIS fusion algorithm, standardize the impact index data in the noise map, calculate the information entropy of each impact index, and allocate the weights of each impact index according to the information entropy; Determine the ideal values of each impact index, and calculate the distance between the corresponding area and the ideal value based on the weights to obtain the noise evaluation result.

8. A quiet community noise map construction and evaluation system based on multi-source data fusion, characterized in that, The system includes: A parameter generation module for: According to the multi-spectral scanning data and the preset acoustic database, perform building surface acoustic heterogeneity modeling processing to obtain a frequency-variable sound absorption parameter matrix; According to the voiceprint recognition result of the sudden noise and the traffic monitoring data, calculate through the sudden noise spatio-temporal association model to obtain the sudden noise impact map; A sound field reconstruction module for performing dynamic sound field reconstruction processing according to the frequency-variable sound absorption parameter matrix combined with real-time sound pressure data to generate a noise heat map; An annoyance quantification module for performing psycho-physiological coupling calculation processing based on the Zwicker loudness model according to the sudden noise impact map combined with electrocardiogram monitoring data, and outputting a composite noise annoyance index; A noise evaluation module for constructing a noise map according to the composite noise annoyance index and the noise heat map, and using the entropy weight-TOPSIS fusion algorithm to perform multi-dimensional index comprehensive evaluation processing through the noise map to obtain a noise evaluation result, where the noise evaluation result is used to generate a noise reduction priority decision scheme.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images