Classroom sound environment optimization design method, device, equipment and medium

Abstract

The invention relates to a classroom sound environment optimization design method and device, equipment and a medium. The method comprises the steps of performing semantic analysis according to a natural language demand text, generating a quantitative acoustic target value set, and performing sound field simulation through a mixed acoustic model in combination with a classroom geometric model; according to the three-dimensional acoustic parameter matrix obtained through simulation and the quantitative acoustic target value set, an acoustic defect reference map is generated; performing spatial constraint clustering analysis on the graph, generating and based on a defect region feature table, performing material selection, calculating a minimum laying area, and generating a material layout scheme; and performing sound field re-simulation processing based on the scheme, calculating a comprehensive deviation value in combination with the quantitative acoustic target value set, and taking the corresponding material layout scheme as a final scheme when the value meets a preset threshold value. According to the method, the accuracy and economical efficiency of classroom sound environment optimization and the reliability of a final scheme are improved by means of natural language demand quantification, defect accurate positioning, material minimization layout and the like.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of computers, and particularly relates to a classroom acoustic environment optimization design method, device, equipment and medium. BACKGROUND

[0002] As a core place for teaching activities, the quality of the acoustic environment of a classroom directly affects the teaching efficiency of teachers and the learning effect of students. Among them, the problems such as too long reverberation time and insufficient language intelligibility will cause the attenuation and distortion of sound signals, and significantly reduce the quality of teaching information transmission. Therefore, the optimization design of the acoustic environment of a classroom is a key link in the design of educational buildings.

[0003] At present, the optimization design of the acoustic environment of a classroom mainly relies on the engineering experience of designers, and mainly determines the type and laying scheme of acoustic materials such as sound absorption and diffusion through subjective judgment, or only uses a single sound field simulation tool for rough simulation. However, this design method has obvious shortcomings. On the one hand, users usually describe the acoustic environment requirements in natural language, such as eliminating the echo on the podium or improving the intelligibility in the back row, but the traditional optimization design method of the acoustic environment of a classroom is difficult to accurately convert such fuzzy requirements into quantitative acoustic targets such as reverberation time and language intelligibility, resulting in a disconnection between the design target and the user's demand, and the optimization direction lacks pertinence. On the other hand, the traditional optimization design method of the acoustic environment of a classroom lacks fine simulation analysis of the sound field of the classroom, and mainly adopts the method of uniformly laying materials in the whole space, without accurately positioning the local acoustic defect area, which not only causes waste of materials and high cost, but also cannot realize efficient optimization of the key defect area, and the optimization effect is difficult to guarantee. In addition, the existing sound field simulation technology has technical bottlenecks, and most of the schemes separately use normal mode theory or ray tracing method to carry out simulation. The normal mode theory has high simulation accuracy in the low frequency band, but has large simulation error in the middle and high frequency bands. The ray tracing method is suitable for simulation in the middle and high frequency bands, but cannot accurately reflect the low frequency sound mode distribution characteristics, resulting in insufficient simulation accuracy of the sound field in the whole frequency band, and the generated acoustic parameters cannot fully reflect the acoustic characteristics of the three-dimensional space of the classroom. SUMMARY

[0004] Therefore, it is necessary to provide a classroom acoustic environment optimization design method, device, equipment and medium aiming at the above technical problems, so as to improve the accuracy and efficiency of the optimization design of the acoustic environment of a classroom, and enhance the pertinence and optimization effect of the optimization design of the acoustic environment of a classroom.

[0005] In a first aspect, the present application provides a classroom acoustic environment optimization design method, comprising:

[0006] performing semantic analysis processing on the natural language requirement text input by the user to generate a set of quantitative acoustic target values, and obtaining a preset classroom geometric model;

[0007] Based on the quantized acoustic target value set and the classroom geometric model, the sound field simulation processing is performed through a hybrid acoustic model to generate a three-dimensional acoustic parameter matrix, and the three-dimensional acoustic parameter matrix and the quantized acoustic target value set are compared point by point to calculate the reverberation time deviation value and the speech intelligibility deviation value, and an acoustic defect benchmark map is generated.

[0008] The acoustic defect benchmark map is subjected to spatial constraint clustering analysis processing to generate a defect area feature table, and based on the defect area feature table, material selection is performed through a material matching rule library to obtain a material selection result, and the minimum laying area is calculated in combination with the defect area feature table to generate a material layout scheme.

[0009] Based on the material layout scheme, sound field re-simulation processing is performed to generate a new acoustic parameter matrix, and the new acoustic parameter matrix and the quantized acoustic target value set are subjected to comprehensive deviation value calculation processing to obtain a comprehensive deviation value, and when the comprehensive deviation value meets a preset threshold value, the corresponding material layout scheme is output as a final scheme.

[0010] In one of the embodiments, the natural language requirement text input by the user is subjected to semantic analysis processing to generate a quantized acoustic target value set, including:

[0011] The natural language requirement text is subjected to keyword extraction processing to generate an acoustic problem type code;

[0012] A preset fuzzy rule library is acquired, and the acoustic problem type code is input into the fuzzy rule library for matching processing to obtain an acoustic parameter adjustment rule;

[0013] Based on the acoustic parameter adjustment rule, quantization conversion processing is performed to generate a quantized acoustic target value set, and the quantized acoustic target value set includes a reverberation time target value and a speech intelligibility target value.

[0014] In one of the embodiments, based on the quantized acoustic target value set and the classroom geometric model, the sound field simulation processing is performed through a hybrid acoustic model to generate a three-dimensional acoustic parameter matrix, including:

[0015] The hybrid acoustic model is called, and the hybrid acoustic model includes a low-frequency calculation module, a medium-high frequency calculation module, and a fusion module, the low-frequency calculation module is constructed by using normal mode theory, and the medium-high frequency calculation module is constructed by using ray tracing method;

[0016] Based on the quantized acoustic target value set, low-frequency band calculation parameters and medium-high frequency band calculation parameters are determined;

[0017] Obtaining a preset frequency boundary value, inputting a classroom geometric model into a low-frequency calculation module of a hybrid acoustic model, combining low-frequency band calculation parameters, and performing acoustic mode distribution calculation and processing on a low-frequency band with a frequency lower than the preset frequency boundary value through normal mode theory to generate a low-frequency band acoustic mode distribution, the low-frequency band acoustic mode distribution containing low-frequency acoustic parameters of each coordinate point in the three-dimensional space of the classroom;

[0018] Inputting the classroom geometric model into a medium-high frequency calculation module of the hybrid acoustic model, combining medium-high frequency band calculation parameters, and performing acoustic energy attenuation distribution calculation and processing on a medium-high frequency band with a frequency not lower than the preset frequency boundary value through ray tracing to generate a medium-high frequency band acoustic energy attenuation distribution, the medium-high frequency band acoustic energy attenuation distribution containing medium-high frequency acoustic parameters of each coordinate point in the three-dimensional space of the classroom;

[0019] Inputting the low-frequency band acoustic mode distribution and the medium-high frequency band acoustic energy attenuation distribution into a fusion module of the hybrid acoustic model, performing fusion processing on the low-frequency acoustic parameters and the medium-high frequency acoustic parameters of the coordinate points in the same three-dimensional space of the classroom, and generating a three-dimensional acoustic parameter matrix, the three-dimensional acoustic parameter matrix containing corresponding comprehensive acoustic parameters of each coordinate point in the three-dimensional space of the classroom.

[0020] In one of the embodiments, the acoustic defect benchmark map is subjected to spatial constraint clustering analysis processing to generate a defect area feature table, including:

[0021] Extracting a reverberation time deviation value and a speech intelligibility deviation value from the acoustic defect benchmark map, and performing screening processing on the reverberation time deviation value and the speech intelligibility deviation value based on a preset deviation threshold to obtain a set of out-of-standard points;

[0022] Obtaining a preset neighborhood radius and a preset minimum point number, and performing spatial constraint clustering processing on the set of out-of-standard points based on the preset neighborhood radius and the preset minimum point number to obtain a plurality of connected defect areas;

[0023] Performing feature extraction processing on each connected defect area to obtain and combine the area identifier, the area vertex coordinates, the area area, the average reverberation time deviation value, and the average speech intelligibility deviation value of each connected defect area to generate a defect area feature table.

[0024] In one of the embodiments, based on the defect area feature table, material selection is performed through a material matching rule library to obtain a material selection result, and the minimum laying area is calculated based on the defect area feature table to generate a material layout scheme, including:

[0025] Obtaining a material matching rule library, extracting an average reverberation time deviation value and an average speech intelligibility deviation value from the defect area feature table, inputting the average reverberation time deviation value and the average speech intelligibility deviation value into the material matching rule library for matching processing to obtain a material selection result, the material selection result including a material model and corresponding material acoustic parameters;

[0026] Extract the material sound absorption coefficient from the material selection results, and extract the area and average reverberation time deviation from the defect area characteristic table;

[0027] The minimum laying area is obtained by calculating based on the material's sound absorption coefficient, the area of ​​the region, and the average reverberation time deviation.

[0028] Extract area identifiers from the defect area feature table, integrate area identifiers, material selection results, and minimum laying area to generate a material layout scheme.

[0029] In one embodiment, a sound field re-simulation is performed based on the material layout scheme to generate a new acoustic parameter matrix. The new acoustic parameter matrix is ​​then compared with the quantized acoustic target value set to calculate a comprehensive deviation value, including:

[0030] The acoustic parameters of the materials are extracted from the material layout scheme and then injected into the classroom geometric model for updating, resulting in the updated classroom geometric model.

[0031] Based on the set of quantized acoustic target values ​​and the updated classroom geometric model, a new acoustic parameter matrix is ​​generated by resimulating the sound field through a hybrid acoustic model.

[0032] Obtain preset weighting coefficients, and perform weighted calculations on the deviations between the new acoustic parameter matrix and the set of quantized acoustic target values ​​based on these preset weighting coefficients to obtain a comprehensive deviation value.

[0033] In one embodiment, the minimum laying area is calculated using the following formula:

[0034]

[0035] in, For the first The minimum laying area for each connected defect region, 0.161 is the Sabine reverberation formula constant. For the first The classroom sub-volume corresponding to each connected defect region. For the first The average reverberation time deviation of each connected defect region This is a region shape correction factor. The number of frequency points for acoustic calculation. The weighting coefficients for frequency f are... This refers to the sound absorption coefficient of the selected material at frequency f, extracted from the material selection results. The sound absorption coefficient of the original material in the connected defect region at frequency f. The average incident angle of the sound waves in the connected defect region. For frequency correction factor, The reference frequency is 1000 Hz.

[0036] In a second aspect, the present application further provides a classroom acoustic environment optimization design device, comprising:

[0037] A demand analysis and model acquisition module is configured to perform semantic analysis processing on a natural language demand text input by a user, generate a set of quantitative acoustic target values, and acquire a preset classroom geometric model.

[0038] An acoustic field simulation and defect analysis module is configured to perform acoustic field simulation processing on the set of quantitative acoustic target values and the classroom geometric model based on a hybrid acoustic model, generate a three-dimensional acoustic parameter matrix, and perform point-by-point comparison processing on the three-dimensional acoustic parameter matrix and the set of quantitative acoustic target values, calculate a reverberation time deviation value and a speech intelligibility deviation value, and generate an acoustic defect benchmark map.

[0039] A material selection and layout generation module is configured to perform spatial constraint clustering analysis processing on the acoustic defect benchmark map, generate a defect area feature table, select materials based on the defect area feature table by using a material matching rule library, obtain a material selection result, and calculate a minimum laying area in combination with the defect area feature table, and generate a material layout scheme.

[0040] A scheme verification and output module is configured to perform acoustic field re-simulation processing based on the material layout scheme, generate a new acoustic parameter matrix, perform comprehensive deviation value calculation processing on the new acoustic parameter matrix and the set of quantitative acoustic target values, obtain a comprehensive deviation value, and output the corresponding material layout scheme as a final scheme when the comprehensive deviation value meets a preset threshold.

[0041] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the first aspect when executing the computer program.

[0042] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps in the first aspect.

[0043] The classroom acoustic environment optimization design method, device, equipment and medium provided by the above classroom acoustic environment optimization design method, device, equipment and medium first perform semantic analysis on the natural language requirement text, obtain a classroom geometric model, accurately convert fuzzy requirements into calculable acoustic parameter targets, and provide clear basis for subsequent design, avoiding the defect that the design target is disengaged from the user requirements due to traditional dependence on experience. Secondly, based on the quantified target and the classroom geometric model, a three-dimensional acoustic parameter matrix is generated through hybrid acoustic model simulation, and an acoustic defect benchmark map is generated by comparing and calculating the deviation, which can realize fine simulation and defect visualization of the full-band sound field, solve the problems of insufficient accuracy and fuzzy defect positioning of a single simulation method. Moreover, a feature table is generated by spatial constraint clustering of the defect benchmark map, and the table is combined to select materials and calculate the minimum laying area to generate a layout scheme, realizing accurate optimization of local defects and making up for the deficiencies of poor pertinence and high cost of existing designs. Finally, a new acoustic parameter matrix is generated by re-simulation based on the layout scheme, the comprehensive deviation value is calculated, and the final scheme is output when the standard is met, effectively solving the risk of no verification and unstable optimization effect of traditional design, and significantly improving the accuracy, economy and reliability of the classroom acoustic environment optimization. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiment or related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0045] Figure 1 A flow chart of a classroom acoustic environment optimization design method is provided for an exemplary embodiment of the present application.

[0046] Figure 2 A flow chart of a method for generating a defect area feature table is provided for an exemplary embodiment of the present application.

[0047] Figure 3 A structural schematic diagram of a classroom acoustic environment optimization design device is provided for an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0049] In one embodiment, as Figure 1As shown, a classroom acoustic environment optimization design method is provided, and in this embodiment, the method is applied to a terminal for illustration. It can be understood that the 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 of the terminal and the server. In this embodiment, the method includes the following steps:

[0050] S101: Perform semantic analysis processing on the natural language requirement text input by the user, generate a set of quantified acoustic target values, and obtain a preset classroom geometric model.

[0051] Specifically, as a non-professional technical personnel, the user expresses the requirements for the classroom acoustic environment in the form of natural language, such as reducing the noise in the back row of the classroom and improving the sound clarity in the podium area. However, this expression lacks quantifiable acoustic parameters, and if it is directly used for design, it will lead to ambiguous optimization direction and deviated effect. Therefore, the natural language processing technology can be used to perform semantic analysis on the natural language requirement text to obtain a set of quantified acoustic target values corresponding to the classroom. The geometric structure of the classroom is the basic carrier of the sound field propagation, and the distribution characteristics of the sound field are directly related to the length, width and height dimensions, wall layout, door and window position and other geometric parameters of the classroom. Lack of accurate geometric model will lead to distortion of the subsequent sound field simulation results. Therefore, a preset classroom geometric model can be obtained first. The model can be derived from building design BIM and includes three-dimensional spatial coordinate information of the classroom, initial acoustic properties of the wall material, position and size data of internal facilities (such as desks, chairs and podiums), etc., providing an accurate spatial basis for subsequent sound field simulation.

[0052] S102: Based on the set of quantified acoustic target values and the classroom geometric model, perform sound field simulation processing through a hybrid acoustic model to generate a three-dimensional acoustic parameter matrix, and perform point-by-point comparison processing on the three-dimensional acoustic parameter matrix and the set of quantified acoustic target values to calculate the reverberation time deviation value and the language clarity deviation value, and generate an acoustic defect benchmark map.

[0053] Specifically, the propagation characteristics of different frequency bands of sound waves differ significantly, among which the wavelength of low-frequency sound waves is long, on the same order of magnitude as the size of the classroom space, and its propagation is significantly affected by the distribution of space modes, making it difficult to accurately simulate using ray tracing alone. The wavelength of mid-high frequency sound waves is shorter, and the propagation path is almost straight, with more obvious effects of reflection and scattering, and the normal wave theory alone cannot accurately reflect the sound energy decay law. Therefore, this step adopts a hybrid acoustic model that combines the advantages of the two simulation methods to achieve accurate simulation of the full-band sound field. By determining the simulation accuracy requirements based on the quantitative acoustic target value set, the classroom geometry model is input into the hybrid acoustic model, and the low-frequency and mid-high frequency bands are simulated respectively. Finally, a three-dimensional acoustic parameter matrix containing the acoustic parameters of each coordinate point in the three-dimensional space of the classroom is output, which records the actual values of the key indicators such as the reverberation time and speech intelligibility of each position in the classroom. In addition, by comparing the three-dimensional acoustic parameter matrix with the quantitative acoustic target value set point by point, the corresponding acoustic parameters of each spatial coordinate point can be calculated respectively, and the reverberation time deviation and speech intelligibility deviation can be obtained. Thus, the difference between the current sound environment and the target requirements can be determined. By associating the deviation values of all coordinate points with the spatial position information, an acoustic defect benchmark map can be generated, which presents the distribution range, position coordinates and deviation severity of the acoustic defects inside the classroom in a visual manner, effectively solving the technical problem of traditional methods that cannot accurately locate local acoustic defects.

[0054] S103: Perform spatial constraint clustering analysis on the acoustic defect benchmark map to generate a defect area feature table. Based on the defect area feature table, select materials through a material matching rule library to obtain a material selection result. Combine the defect area feature table with the material selection result to calculate the minimum laying area and generate a material layout scheme.

[0055] Specifically, the exceeding points recorded in the acoustic defect benchmark map may be scattered and distributed, and if each scattered point is processed individually, the construction complexity and material usage will be increased. Therefore, based on the spatial coherence characteristics of sound wave propagation, reasonable spatial constraint conditions can be set, and the spatial constraint clustering analysis can be used to aggregate the exceeding points with spatial continuity into connected defect regions. Since the regions formed by clustering are the ranges of acoustic defects that are concentrated and have physical relevance, the core features of each region can be extracted, including the region vertex coordinates, the region area, the average reverberation time deviation value, and the average speech intelligibility deviation value, to generate a defect region feature table. Different acoustic defects correspond to different material functional requirements, for example, when the reverberation time deviation value is too large, a material with strong sound absorption performance needs to be selected, and when the speech intelligibility deviation value is insufficient, a material with diffusion performance needs to be selected. The material matching rule library is constructed based on industry technical standards and engineering practice data, and the appropriate material can be automatically selected according to the deviation type and value in the defect region feature table to form a material selection result. Based on the result, the minimum laying area can be calculated by comprehensively considering the defect region area, the average deviation value, and the acoustic performance parameters of the selected material. In combination with the material selection result and the minimum laying area integration, a material layout scheme can be generated, which not only ensures the effective optimization of the defect region, but also avoids the waste of resources caused by global laying, achieving a balance between optimization effect and economy.

[0056] S104: Based on the material layout scheme, a sound field re-simulation process is performed to generate a new acoustic parameter matrix, a comprehensive deviation value is calculated by comprehensively processing the new acoustic parameter matrix and the quantized acoustic target value set, and when the comprehensive deviation value meets a preset threshold, the corresponding material layout scheme is output as a final scheme.

[0057] Specifically, the material layout scheme is only a theoretical design, and its actual application effect needs to be verified through sound field re-simulation. The technical principle of this process is consistent with that of the sound field simulation in S102, and the difference lies in that the material acoustic parameters in the material layout scheme need to be injected into the classroom geometry model to update the acoustic properties of the model, and then the full-band sound field simulation is performed again through the hybrid acoustic model to generate a new acoustic parameter matrix, which reflects the expected sound environment state of the classroom after implementing the material layout scheme. And the comprehensive deviation value can be calculated according to the deviation of each acoustic parameter in the new acoustic parameter matrix and the quantized acoustic target value set. The preset threshold is a pass line set according to the acoustic performance standard of the classroom sound environment, representing the maximum acceptable deviation range. When the comprehensive deviation value is less than the threshold, it means that the material layout scheme can make the classroom sound environment meet the user's demand, and it can be output as the final scheme, thereby ensuring the effectiveness and reliability of the final output scheme and avoiding the problem of engineering rework or substandard optimization due to design omissions.

[0058] The above method firstly solves the problem of disconnection between user demand and design target in traditional methods by performing semantic analysis on the natural language demand text input by the user. Secondly, based on the set of quantitative acoustic target values and the classroom geometric model, a three-dimensional acoustic parameter matrix is generated through sound field simulation and the deviation value is calculated to generate an acoustic defect benchmark map, solving the problem of lack of fine simulation analysis in traditional methods and improving the accuracy of acoustic defect positioning. Moreover, through spatial constraint clustering analysis of the acoustic defect benchmark map, a defect area feature table is generated, and based on the feature table, material selection and layout scheme generation are performed, solving the problem of excessive material usage, high cost and poor optimization effect in traditional methods, and realizing precise material laying and efficient optimization. Finally, based on the material layout scheme, sound field re-simulation is performed to calculate the comprehensive deviation value, and when the preset threshold is met, the final scheme is output, solving the problem of lack of targeted optimization direction and insufficient full-band sound field simulation accuracy in traditional methods, and improving the accuracy and effect of classroom sound environment optimization design.

[0059] In one embodiment, the natural language demand text input by the user is processed for semantic analysis to generate a set of quantitative acoustic target values, including:

[0060] The natural language demand text is processed for keyword extraction to generate an acoustic problem type code;

[0061] A preset fuzzy rule library is obtained, and the acoustic problem type code is input into the fuzzy rule library for matching processing to obtain an acoustic parameter adjustment rule;

[0062] Based on the acoustic parameter adjustment rule, a quantitative conversion process is performed to generate a set of quantitative acoustic target values, including a reverberation time target value and a speech intelligibility target value.

[0063] Specifically, based on the semantic focus and field vocabulary matching technology in natural language processing, a special vocabulary library for the acoustic field can be constructed, which includes core acoustic concepts such as reverberation, echo, intelligibility, noise, volume, and corresponding synonymous expressions. By using a pre-trained semantic understanding model such as the BERT model, the input natural language requirement text can be segmented, tagged with parts of speech, and analyzed for semantic association. The vocabulary with a matching degree higher than a preset threshold with the acoustic field vocabulary library can be selected as the core keyword. For example, for the requirement text "hope to reduce the echo in the corner of the classroom and make the middle area clearer", the extracted keywords are echo and intelligibility. Then, based on the preset encoding rules, the extracted keywords are standardized and converted. Each acoustic problem type corresponds to a unique encoding identifier, for example, echo corresponds to encoding P1, and insufficient intelligibility corresponds to encoding P2. Through this encoding conversion, the natural language keywords can be converted into standardized acoustic problem type encodings that are easy to match with the rule base, avoiding the ambiguity of natural language expression and laying the foundation for accurate mapping of acoustic parameters.

[0064] Further, since there is no absolute rigid mapping relationship between the acoustic problems corresponding to the user requirements and the quantitative parameters, it is necessary to combine engineering practice experience and industry standards to construct the corresponding matching logic. Illustratively, a preset fuzzy rule base can be obtained, which can be constructed based on the preset industry standards and combined with a large number of classroom acoustic environment optimization engineering case data, and each rule in the rule base includes a precondition (acoustic problem type encoding), a post-conclusion (acoustic parameter adjustment rule), and a confidence threshold. In the matching process, the generated acoustic problem type encoding is input into the fuzzy rule base, and all rules in the rule base can be traversed by a forward reasoning algorithm to filter out rules with completely matched preconditions and confidence higher than the preset threshold. The post-conclusion of the rule is integrated into the acoustic parameter adjustment rule, which clearly defines the type of acoustic parameter to be optimized, the target range, and the adjustment priority, solving the adaptation fault problem between natural language requirements and acoustic technical parameters.

[0065] Specifically, due to the differences in space size and use scenarios (such as primary school classrooms and university auditoriums) of different classrooms, the optimal quantization target values corresponding to the same acoustic problem will also be different, and therefore, the quantization conversion processing based on the acoustic parameter adjustment rule can be accurately calibrated in combination with the core features of the classroom scenario. Illustratively, key scene parameters can be extracted from the obtained classroom geometric model, including the classroom volume, the length-width-height ratio, the maximum distance between the podium and the back row, and the like. These scene parameters and the target range in the acoustic parameter adjustment rule are substituted into a preset quantization calculation model, which is constructed based on acoustic theory formulas and engineering experience coefficients. For example, for the quantization calculation of the reverberation time, in combination with the classroom volume V, the specific value of the reverberation time T that satisfies the target range can be inversely calculated through the formula T = 0.16V / A (where A is the total sound absorption amount), and if the reverberation time T target range given by the rule is 0.4-0.6s, the final quantization value is determined to be 0.5s after calculation in combination with the classroom volume V. For the language intelligibility target value, the corresponding quantization value can be calculated in combination with the air sound attenuation coefficient according to the maximum distance between the back row and the acoustic parameter adjustment rule minimum requirement (such as greater than 0.75). Through this quantization conversion process, the target range in the acoustic parameter adjustment rule can be converted into a unique and accurate quantization value, generating a quantized acoustic target value set containing the reverberation time target value and the language intelligibility target value. This set provides a clear quantization benchmark for subsequent sound field simulation, deviation calculation and material optimization, ensuring the pertinence and operability of the entire optimization process and avoiding design direction deviation caused by ambiguous target parameters.

[0066] In one embodiment, based on the quantized acoustic target value set and the classroom geometric model, a sound field simulation process is performed through a hybrid acoustic model to generate a three-dimensional acoustic parameter matrix, including:

[0067] The hybrid acoustic model includes a low-frequency calculation module, a medium-high frequency calculation module and a fusion module, the low-frequency calculation module is constructed by using the normal mode theory, and the medium-high frequency calculation module is constructed by using the ray tracing method;

[0068] Based on the quantized acoustic target value set, low-frequency band calculation parameters and medium-high frequency band calculation parameters are determined;

[0069] A preset frequency division value is obtained, the classroom geometric model is input into the low-frequency calculation module of the hybrid acoustic model, and the low-frequency band sound mode distribution calculation process is performed on the low-frequency band with a frequency lower than the preset frequency division value through the normal mode theory in combination with the low-frequency band calculation parameters, to generate a low-frequency band sound mode distribution, the low-frequency band sound mode distribution containing low-frequency acoustic parameters of each coordinate point in the three-dimensional space of the classroom;

[0070] The classroom geometry model is input into a mid-high frequency calculation module of the hybrid acoustic model, combined with mid-high frequency band calculation parameters, and the mid-high frequency band with a frequency not lower than a preset frequency demarcation value is calculated and processed by a ray tracing method to generate a mid-high frequency sound energy attenuation distribution, and the mid-high frequency sound energy attenuation distribution contains mid-high frequency acoustic parameters of each coordinate point in the three-dimensional space of the classroom.

[0071] The low-frequency sound mode distribution and the mid-high frequency sound energy attenuation distribution are input into a fusion module of the hybrid acoustic model, and the low-frequency acoustic parameters and the mid-high frequency acoustic parameters of the coordinate points in the same three-dimensional space of the classroom are fused to generate a three-dimensional acoustic parameter matrix, and the three-dimensional acoustic parameter matrix contains corresponding comprehensive acoustic parameters of each coordinate point in the three-dimensional space of the classroom.

[0072] Specifically, the hybrid acoustic model can integrate the simulation logic of low frequency and mid-high frequency bands through modular design. Among them, the low-frequency sound wave has a long wavelength (usually greater than 1 / 4 of the size of the classroom space), its propagation is significantly affected by the boundary conditions of the classroom, and the sound energy distribution presents obvious modal characteristics, while the normal mode theory can accurately describe the spatial distribution of sound pressure level in the low frequency band by solving the eigenvalues and eigenfunctions of the sound wave equation under specific boundary conditions, so the low frequency calculation module can be constructed by using the normal mode theory. The mid-high frequency sound wave has a short wavelength, and the propagation path is approximately straight, and the energy attenuation is mainly achieved by reflection, scattering and air absorption, while the ray tracing method can effectively reflect the attenuation law of mid-high frequency sound energy by simulating the propagation path and energy loss of sound rays, so the mid-high frequency calculation module is constructed by using the ray tracing method. The fusion module is responsible for spatiotemporal alignment and weight fusion of the simulation results of the two frequency bands to ensure the continuity and consistency of the output parameters. Through the above modular structure of the hybrid acoustic model, the simulation advantages of different theories in the corresponding frequency band are retained, and the integration of full-band parameters is realized through a unified framework, solving the precision short board of single simulation method in full-band coverage.

[0073] Specifically, the reverberation time, speech intelligibility and other parameters in the set of quantized acoustic target values are closely related to specific frequency bands. For example, the reverberation time is more affected by modal superposition in the low frequency band, and the speech intelligibility mainly depends on the sound energy distribution in the medium and high frequency band. Therefore, the frequency band related constraint conditions can be extracted from the quantized target as the calculation parameters. For example, the target attenuation coefficients of low frequency bands such as 125Hz and 250Hz are extracted from the reverberation time target value, and the target sound pressure level range of medium and high frequency bands such as 500Hz, 1000Hz and 2000Hz is extracted from the speech intelligibility target value. These parameters are input into the low frequency calculation module and the medium and high frequency calculation module as the convergence conditions and precision control indicators of the simulation process, to ensure that the simulation results directly serve the subsequent deviation calculation from the target value, and improve the pertinence and efficiency of the simulation. Then, a preset frequency demarcation value can be obtained and simulated in different frequency bands. The frequency demarcation value can be set based on the turning point of the sound wave propagation characteristics, for example, the division of low frequency and medium and high frequency in the classroom acoustic design, such as 200Hz or 315Hz, and the value needs to ensure that the simulation error of the normal mode theory and the ray tracing method in each frequency band is less than a preset error threshold, such as 5%.

[0074] Specifically, by inputting the classroom geometric model into the low frequency calculation module, the three-dimensional coordinate system can be constructed according to the classroom geometric model, and the material acoustic properties (such as sound absorption coefficient, reflection coefficient) of the walls, floor and ceiling are defined as boundary conditions. Then, by solving the normal frequency and normal mode of Helmholtz equation ∇²p+k²p=0 (where p is the sound pressure and k is the wave number) under the boundary conditions, the sound pressure superposition value of each normal mode at each coordinate point in the three-dimensional space of the classroom can be calculated, and the low frequency band sound mode distribution including sound pressure level, modal density and other parameters can be obtained. The parameters of each coordinate point correspond to multiple characteristic frequencies in the low frequency band, such as 125Hz and 250Hz. For the medium and high frequency calculation module, the classroom geometric model and the medium and high frequency calculation parameters are input, and sound rays are emitted at a preset angle interval, such as 1°, at the sound source position (such as the podium area), each sound ray carrying initial sound energy. Then, according to the wall and object surface properties in the classroom geometric model, the reflection times, reflection angles and energy attenuation (including air absorption and material absorption loss) of the sound rays during propagation are calculated. When the sound ray energy is below a preset threshold or the propagation distance exceeds the classroom size, the tracking is terminated, and by counting the total energy and arrival time distribution of the sound rays received at each coordinate point in the three-dimensional space of the classroom, a medium and high frequency band sound energy decay distribution including sound pressure level, early decay time and other parameters can be generated, covering medium and high frequency characteristic frequencies such as 500Hz to 4000Hz. The above frequency band simulation significantly improves the calculation accuracy of the acoustic parameters in each frequency band by adapting the physical characteristics of different frequency bands, providing high-quality basic data for subsequent fusion processing.

[0075] Specifically, the simulation results of the low-frequency and mid-to-high-frequency bands are input into the fusion module for processing, which generates a three-dimensional acoustic parameter matrix covering the entire frequency band, enabling a comprehensive description of the classroom's acoustic environment. For frequency bands below the boundary value, the parameters of the low-frequency acoustic mode distribution are primarily used, with mid-to-high-frequency parameters assigned lower weights (e.g., 0.1-0.3) based on the frequency attenuation coefficient. For frequency bands above the boundary value, the parameters of the mid-to-high-frequency acoustic energy attenuation distribution are used primarily, with the weight of low-frequency parameters gradually decreasing as the frequency increases (e.g., from 0.3 to 0). In transitional frequency bands near the boundary value, such as ±1 / 3 octave band of the boundary value, a linear weighting method can be used to fuse the two parameter segments, avoiding abrupt changes. In a schematic manner, acoustic parameters at characteristic frequencies in both the low-frequency and mid-to-high-frequency bands are extracted for each coordinate point within the three-dimensional space of the classroom. These parameters are then fused according to the aforementioned weighting rules. For example, the sound pressure level (SPL) at 200Hz (the dividing line) is obtained by superimposing the 200Hz low-frequency parameter (weight 0.5) and the 200Hz mid-to-high-frequency parameter (weight 0.5). The SPL at 1000Hz is obtained by superimposing the 1000Hz mid-to-high-frequency parameter (weight 0.9) and the 1000Hz low-frequency parameter (weight 0.1). This fusion process generates a three-dimensional acoustic parameter matrix containing comprehensive acoustic parameters for each coordinate point in the three-dimensional space of the classroom across the entire frequency range, from 125Hz to 4000Hz. These comprehensive acoustic parameters can include SPL, reverberation time, and speech intelligibility index. This matrix retains the precise characteristics of each frequency band while achieving continuous and unified parameters across the entire frequency range, providing a complete and reliable sound field data foundation for subsequent point-by-point comparisons with quantified acoustic target values.

[0076] In one embodiment, such as Figure 2 As shown, spatially constrained clustering analysis is performed on the acoustic defect baseline map to generate a defect region feature table, including:

[0077] S201: Extract reverberation time deviation and speech intelligibility deviation values ​​from the acoustic defect reference map, and filter the reverberation time deviation and speech intelligibility deviation values ​​based on a preset deviation threshold to obtain a set of out-of-standard points;

[0078] S202: Obtain the preset neighborhood radius and preset minimum number of points, and perform spatial constraint clustering on the set of out-of-standard points based on the preset neighborhood radius and preset minimum number of points to obtain multiple connectivity defect regions;

[0079] S203: Perform feature extraction processing on each connected defect region to obtain and combine the region identifier, region vertex coordinates, region area, average reverberation time deviation value and average speech intelligibility deviation value of each connected defect region to generate a defect region feature table.

[0080] Specifically, preset deviation thresholds can be determined based on the allowable fluctuation range in the set of quantitative acoustic target values, combined with the functional requirements of the classroom acoustic environment. For example, the reverberation time deviation threshold can be set to ±10% of the target value, and the speech intelligibility deviation threshold can be set to ±5% of the target value. By traversing all three-dimensional spatial coordinate points in the acoustic defect benchmark map, and comparing the reverberation time deviation value of each point with the preset reverberation deviation threshold or the speech intelligibility deviation value with the preset intelligibility deviation threshold, coordinate points that meet at least one of the exceeding conditions can be included in the set, forming an exceeding point set. This ensures that subsequent cluster analysis only targets the areas that truly need optimization, avoiding invalid calculations and wasted resources.

[0081] Furthermore, since the generation and propagation of acoustic defects are spatially correlated—for example, excessive reverberation caused by wall reflections is usually concentrated in specific areas rather than isolated points—clustering needs to introduce spatial constraints. The preset neighborhood radius is related to the sound wave wavelength and can be set to half the longest wavelength within the optimized frequency band. For example, for a 200Hz sound wave with a wavelength of approximately 1.7m, the neighborhood radius can be set to 0.85m, ensuring that physically acoustically correlated points are grouped into the same area. The preset minimum number of points can be determined based on the minimum effective size of the defect area, for example, set to 5 points, to avoid misclassifying a small number of discrete points as effective defect areas. Subsequently, a density clustering algorithm such as DBSCAN can be used for clustering. Unclassified points in the excess point set are randomly selected, and all surrounding excess points are searched using the preset neighborhood radius. If the number of points within this range is not less than the preset minimum number of points, these points are marked as the same connected defect area, and the search for neighboring points of other points within this area is recursively performed until no new points are added. Repeating the above process for the remaining unclassified points yields multiple independent connected defect regions. This clustering method ensures the physical rationality of the regions through spatial constraints, solving the fragmentation problem caused by traditional unconstrained clustering and making the obtained defect regions more closely match the distribution characteristics of actual acoustic defects.

[0082] Specifically, feature extraction is performed on each connected defect region to generate a defect region feature table. This transforms the regionalized defect information into structured data that can be directly used for material selection and area calculation. Feature extraction must cover the spatial location, size, and acoustic deviation of the region. The region identifier is a unique code for each connected defect region, used for subsequent material layout and region mapping. The region vertex coordinates can be determined using the convex hull algorithm to calculate the minimum bounding polygon vertices of the region, accurately describing its spatial location and shape. The region area is calculated based on the polygon area using vertex coordinates, reflecting the coverage of the defect region. The average reverberation time deviation and average speech intelligibility deviation can be obtained by taking the arithmetic mean of the corresponding deviation values ​​of all points within the region, quantifying the severity of the defect. By extracting these features, the defect status of the region can be understood. Integrating the feature information of all connected defect regions forms a defect region feature table. This table provides precise input parameters for subsequent material matching rule base calls, ensuring that material selection and laying area calculation closely match the defect characteristics of each region, avoiding the insufficient specificity of traditional globally unified solutions.

[0083] In one embodiment, based on a defect area feature table, materials are selected using a material matching rule base to obtain material selection results. Combined with the defect area feature table, the minimum laying area is calculated to generate a material layout scheme, including:

[0084] Obtain the material matching rule base, extract the average reverberation time deviation value and the average speech intelligibility deviation value from the defect area feature table, input the average reverberation time deviation value and the average speech intelligibility deviation value into the material matching rule base for matching processing, and obtain the material selection result, which includes the material model and the corresponding material acoustic parameters;

[0085] Extract the material sound absorption coefficient from the material selection results, and extract the area and average reverberation time deviation from the defect area characteristic table;

[0086] The minimum laying area is obtained by calculating based on the material's sound absorption coefficient, the area of ​​the region, and the average reverberation time deviation.

[0087] Extract area identifiers from the defect area feature table, integrate area identifiers, material selection results, and minimum laying area to generate a material layout scheme.

[0088] Specifically, the material matching rule base can include various acoustic materials such as sound-absorbing materials and diffusion materials. Each rule includes triggering conditions (ranges of average reverberation time deviation and average speech intelligibility deviation), material type, and corresponding acoustic parameters such as absorption coefficient frequency curve and scattering coefficient. For example, when the average reverberation time deviation is greater than 0.1s and the average speech intelligibility deviation is within a reasonable range, the rule base matches polyester fiber sound-absorbing panels with an absorption coefficient α > 0.8. When the average speech intelligibility deviation is less than -0.05 and the reverberation time deviation is within a reasonable range, diffuser materials with a scattering coefficient s > 0.7 can be matched. Illustratively, after extracting the average reverberation time deviation and average speech intelligibility deviation from the defect area feature table, and inputting them into the material matching rule base, the system can filter out material types and corresponding acoustic parameters that meet the defect repair requirements through rule traversal and condition matching, forming the material selection results. This process, by replacing manual experience judgment with standardized rules, further ensures the scientific and targeted nature of material selection and avoids optimization failures caused by material performance mismatches.

[0089] Specifically, the sound absorption coefficient of a material is a key indicator reflecting its acoustic performance, directly determining its ability to absorb sound energy. The area reflects the spatial extent of the defective region and is the basis for calculating the amount of material needed. The average reverberation time deviation quantifies the degree of acoustic deviation that needs to be eliminated in that area. Then, based on the material's sound absorption coefficient, area, and average reverberation time deviation, the minimum laying area can be calculated by inversely deriving the Sabine reverberation time formula. The specific calculation formula is as follows:

[0090]

[0091] in, For the first The minimum laying area for each connected defect region, 0.161 is the Sabine reverberation formula constant. For the first The classroom sub-volume corresponding to each connected defect region. For the first The average reverberation time deviation of each connected defect region This is a region shape correction factor. The number of frequency points for acoustic calculation. The weighting coefficients for frequency f are... This refers to the sound absorption coefficient of the selected material at frequency f, extracted from the material selection results. The sound absorption coefficient of the original material in the connected defect region at frequency f. The average incident angle of the sound waves in the connected defect region. For frequency correction factor, This is the reference frequency.

[0092] In the above formula, the classroom sub-volume It can be calculated from the area and height of the defective region (extracted from the classroom geometry model), while the region shape correction coefficient... The value can be set according to the spatial shape of the defect area, such as rectangular or irregular shapes. For regular shapes, a value of 1.0 can be used, while for irregular shapes, a value between 1.1 and 1.2 can be used. Furthermore, the number of frequency points for acoustic calculations can be set. Five characteristic frequencies can be selected, such as 125Hz, 250Hz, 500Hz, 1000Hz, and 2000Hz. This formula, through quantitative acoustic theory, correlates material properties, defect levels, and spatial volume, ensuring that the calculation of the minimum laying area meets acoustic standards while avoiding excessive material use, thus achieving a balance between economy and optimization effect.

[0093] Specifically, each defective area is assigned a unique identification code, ensuring a one-to-one correspondence between material layout and defective areas. The material selection results clearly specify the material models used, and the minimum laying area quantifies the material usage. By integrating this information, the generated material layout plan can include the location, material model, material usage, and laying range of each defective area. This plan provides precise technical guidance for construction, ensuring that material laying covers defective areas without creating redundancy. It solves the problems of vague material layout and large construction errors in traditional designs, achieving precise optimization of the classroom acoustic environment.

[0094] In one embodiment, a sound field re-simulation is performed based on a material layout scheme to generate a new acoustic parameter matrix. The new acoustic parameter matrix is ​​then compared with a set of quantized acoustic target values ​​to calculate a comprehensive deviation value, including:

[0095] The acoustic parameters of the materials are extracted from the material layout scheme and then injected into the classroom geometric model for updating, resulting in the updated classroom geometric model.

[0096] Based on the set of quantized acoustic target values ​​and the updated classroom geometric model, a new acoustic parameter matrix is ​​generated by resimulating the sound field through a hybrid acoustic model.

[0097] Obtain preset weighting coefficients, and perform weighted calculations on the deviations between the new acoustic parameter matrix and the set of quantized acoustic target values ​​based on these preset weighting coefficients to obtain a comprehensive deviation value.

[0098] Specifically, the material acoustic parameters included in the material layout scheme, such as the frequency curves of absorption coefficient and scattering coefficient, are quantitative representations of the material's acoustic performance and directly determine the acoustic characteristics of sound field propagation. The classroom geometric model serves as the spatial carrier for sound field simulation; the simulation results are only valuable when the material properties in the model match the actual layout. Illustratively, by traversing the material selection results for each defective area in the material layout scheme, the acoustic parameters of the corresponding materials, such as the absorption coefficient sequence of a certain sound-absorbing material from 125Hz to 4000Hz, are extracted. The placement position of this material in the classroom geometric model is then located (determined by the coordinates of the region's vertices). These acoustic parameters are injected into the corresponding geometric model units, replacing the original material's acoustic properties, resulting in an updated classroom geometric model. Subsequently, a sound field re-simulation can be performed based on the quantified acoustic target value set and the updated classroom geometric model. The technical principle is consistent with the initial sound field simulation, i.e., a hybrid acoustic model is used to simulate the low-frequency and mid-to-high-frequency sound fields of the updated classroom geometric model. For example, in low-frequency simulations, the Helmholtz equations eigenvalues ​​and eigenfunctions, incorporating the acoustic properties of the new material, are re-solved using normal mode theory to calculate the sound pressure distribution of each normal mode in three-dimensional space. In mid-to-high-frequency simulations, ray tracing is used to re-simulate the reflection and absorption of sound rays on the surface of the new material, and the sound energy distribution at each coordinate point is statistically analyzed. Finally, by fusing the simulation results from the low-frequency and mid-to-high-frequency ranges, a new acoustic parameter matrix can be generated. This matrix contains comprehensive acoustic parameters such as reverberation time and speech intelligibility index at each coordinate point in the three-dimensional space of the classroom across the entire frequency range. The changes in these values ​​directly reflect the optimization effect of the material layout on the sound field. For example, if the material layout scheme is effective, the reverberation time deviation value in the defective areas of the new acoustic parameter matrix should be significantly reduced, and the speech intelligibility deviation value should be significantly improved.

[0099] Furthermore, different acoustic parameters have varying degrees of impact on classroom acoustic environment quality. For example, excessively long reverberation time can lead to muddy sound, and insufficient speech intelligibility can affect information transmission. Therefore, weighting coefficients can be used to reflect the importance of each parameter. Preset weighting coefficients can be set based on the functional requirements of the classroom acoustic environment and acoustic evaluation standards. For example, the weight of reverberation time can be set to 0.5, and the weight of speech intelligibility can be set to 0.5, or adjusted according to specific scenarios, such as increasing the weight of speech intelligibility in a multimedia classroom. Then, the reverberation time deviation and speech intelligibility deviation values ​​at each coordinate point in the new acoustic parameter matrix can be subtracted from the corresponding target values ​​in the quantified acoustic target value set to obtain the individual deviation value for each point. The individual deviation values ​​at each point are then weighted and summed according to the preset weighting coefficients to obtain the comprehensive deviation value for each point. Finally, the arithmetic mean of the comprehensive deviation values ​​for all coordinate points or a weighted average based on spatial importance (such as the podium area and the student area) can be taken to obtain the comprehensive deviation value for the entire classroom. This calculation process achieves a comprehensive evaluation of multiple parameters through weighted integration, which avoids the one-sidedness of single-parameter evaluation and ensures that the final solution meets the overall acoustic performance requirements, providing a quantitative basis for determining whether the solution meets the standards.

[0100] Based on the same inventive concept, such as Figure 3 As shown, this application also provides a classroom acoustic environment optimization design device 300 for implementing the classroom acoustic environment optimization design method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more classroom acoustic environment optimization design device embodiments provided below can be found in the limitations of the method embodiments above, and will not be repeated here. The device includes:

[0101] The requirement analysis and model acquisition module 301 is used to perform semantic analysis on the natural language requirement text input by the user, generate a set of quantified acoustic target values, and acquire a preset classroom geometric model.

[0102] The sound field simulation and defect analysis module 302 is used to perform sound field simulation processing based on the set of quantized acoustic target values ​​and the classroom geometric model through a hybrid acoustic model, generate a three-dimensional acoustic parameter matrix, and perform point-by-point comparison processing between the three-dimensional acoustic parameter matrix and the set of quantized acoustic target values ​​to calculate the reverberation time deviation value and speech intelligibility deviation value, and generate an acoustic defect benchmark map.

[0103] The material selection and layout generation module 303 is used to perform spatial constraint clustering analysis on the acoustic defect baseline map, generate a defect area feature table, select materials based on the defect area feature table through the material matching rule library, obtain the material selection result, and calculate the minimum laying area in combination with the defect area feature table to generate a material layout scheme.

[0104] The scheme verification and output module 304 is used to perform sound field re-simulation processing based on the material layout scheme, generate a new acoustic parameter matrix, calculate the comprehensive deviation value by combining the new acoustic parameter matrix with the set of quantized acoustic target values, and output the corresponding material layout scheme as the final scheme when the comprehensive deviation value meets the preset threshold.

[0105] In one exemplary embodiment, the present invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the classroom acoustic environment optimization design method of this application. A multi-core processor is preferred to improve the system's parallel processing capability. The memory provides sufficient temporary storage space to support program execution and data processing. The memory capacity should be large enough to accommodate large amounts of data and computational tasks.

[0106] In one exemplary embodiment, the present invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the classroom acoustic environment optimization design method of the present application. The computer-readable storage medium may include: a read-only memory, a random access memory (RAM), a solid-state drive (SSD), or an optical disc, etc.

[0107] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for optimizing the acoustic environment of a classroom, characterized in that, The method comprises: performing semantic analysis processing on a natural language requirement text input by a user, generating a quantized acoustic target value set, and obtaining a preset classroom geometric model; based on the quantized acoustic target value set and the classroom geometric model, performing sound field simulation processing through a hybrid acoustic model to generate a three-dimensional acoustic parameter matrix, and performing point-by-point comparison processing on the three-dimensional acoustic parameter matrix and the quantized acoustic target value set to calculate a reverberation time deviation value and a speech intelligibility deviation value, and generating an acoustic defect benchmark graph; performing spatial constraint clustering analysis processing on the acoustic defect benchmark graph to generate a defect area feature table, selecting materials based on the defect area feature table through a material matching rule library to obtain a material selection result, and combining the defect area feature table to calculate a minimum laying area to generate a material layout scheme; based on the material layout scheme, performing sound field re-simulation processing to generate a new acoustic parameter matrix, and performing comprehensive deviation value calculation processing on the new acoustic parameter matrix and the quantized acoustic target value set to obtain a comprehensive deviation value, and when the comprehensive deviation value meets a preset threshold, outputting the corresponding material layout scheme as a final scheme.

2. The method of claim 1, wherein, The semantic analysis processing on the natural language requirement text input by the user to generate the quantized acoustic target value set comprises: performing keyword extraction processing on the natural language requirement text to generate an acoustic problem type code; obtain a preset fuzzy rule library, input the acoustic problem type code into the fuzzy rule library for matching processing to obtain an acoustic parameter adjustment rule; based on the acoustic parameter adjustment rule, perform quantization conversion processing to generate the quantized acoustic target value set, the quantized acoustic target value set comprising a reverberation time target value and a speech intelligibility target value.

3. The method of claim 1, wherein, The sound field simulation processing based on the quantized acoustic target value set and the classroom geometric model through the hybrid acoustic model to generate the three-dimensional acoustic parameter matrix comprises: calling the hybrid acoustic model, the hybrid acoustic model comprising a low-frequency calculation module, a medium-high frequency calculation module and a fusion module, the low-frequency calculation module being constructed by normal mode theory, and the medium-high frequency calculation module being constructed by ray tracing method; determine the low-frequency band calculation parameters and the medium-high frequency band calculation parameters based on the quantized acoustic target value set; obtain a preset frequency boundary value, input the classroom geometric model into the low-frequency calculation module of the hybrid acoustic model, combine the low-frequency band calculation parameters, and perform acoustic mode distribution calculation processing on the low-frequency band with a frequency lower than the preset frequency boundary value through the normal mode theory to generate a low-frequency band acoustic mode distribution, the low-frequency band acoustic mode distribution containing low-frequency acoustic parameters of each coordinate point in the classroom three-dimensional space; input the classroom geometry model into the mid-high frequency calculation module of the hybrid acoustic model, combine the mid-high frequency band calculation parameters, and perform acoustic energy attenuation distribution calculation and processing on the mid-high frequency band not lower than the preset frequency demarcation value through the ray tracing method to generate a mid-high frequency band acoustic energy attenuation distribution, wherein the mid-high frequency band acoustic energy attenuation distribution contains the mid-high frequency acoustic parameters of each coordinate point in the three-dimensional space of the classroom; input the low frequency band acoustic mode distribution and the mid-high frequency band acoustic energy attenuation distribution into the fusion module of the hybrid acoustic model, and perform fusion processing on the low frequency acoustic parameters and the mid-high frequency acoustic parameters of the coordinate points in the same three-dimensional space of the classroom to generate a three-dimensional acoustic parameter matrix containing the corresponding comprehensive acoustic parameters of each coordinate point in the three-dimensional space of the classroom.

4. The method of claim 1, wherein, The spatial constraint clustering analysis processing of the acoustic defect benchmark map generates a defect area feature table, including: extract the reverb time deviation value and the speech intelligibility deviation value from the acoustic defect benchmark map, filter the reverb time deviation value and the speech intelligibility deviation value based on a preset deviation threshold to obtain a set of out-of-specification points; obtain a preset neighborhood radius and a preset minimum point number, and perform spatial constraint clustering processing on the set of out-of-specification points based on the preset neighborhood radius and the preset minimum point number to obtain a plurality of connected defect areas; perform feature extraction processing on each connected defect area to obtain and combine the area identifier, area vertex coordinates, area area, average reverb time deviation value, and average speech intelligibility deviation value of each connected defect area to generate the defect area feature table.

5. The method of claim 4, wherein, Based on the defect area feature table, material selection is performed through a material matching rule library to obtain a material selection result, and a material layout scheme is generated by calculating the minimum laying area based on the defect area feature table, including: obtain the material matching rule library, extract the average reverb time deviation value and the average speech intelligibility deviation value from the defect area feature table, input the average reverb time deviation value and the average speech intelligibility deviation value into the material matching rule library for matching processing to obtain the material selection result, wherein the material selection result includes the material model and the corresponding material acoustic parameters; extract the material absorption coefficient from the material selection result, and extract the area area and the average reverb time deviation value from the defect area feature table; based on the material absorption coefficient, the area area, and the average reverb time deviation value, perform calculation processing to obtain the minimum laying area; extract the area identifier from the defect area feature table, integrate the area identifier, the material selection result, and the minimum laying area to generate the material layout scheme.

6. The method of claim 1, wherein, Based on the material layout scheme, perform sound field re-simulation processing to generate a new acoustic parameter matrix, and perform comprehensive deviation value calculation processing on the new acoustic parameter matrix and the quantized acoustic target value set to obtain a comprehensive deviation value, including: extracting a material acoustic parameter from the material layout scheme, injecting the material acoustic parameter into the classroom geometric model for update processing to obtain an updated classroom geometric model; based on the set of quantized acoustic target values and the updated classroom geometric model, performing sound field re-simulation processing through the hybrid acoustic model to generate the new acoustic parameter matrix; obtaining a preset weight coefficient, and performing weighted calculation processing on the deviation of the new acoustic parameter matrix and the set of quantized acoustic target values based on the preset weight coefficient to obtain the comprehensive deviation value.

7. The method of claim 5, wherein, The minimum laying area is calculated by the following formula: wherein, is the minimum laying area of the first th connected defect region, 0.161 is the Sabine reverberation formula constant, is the minimum laying area of the first th connected defect region, is the minimum laying area of the first th connected defect region, is the area shape correction coefficient, is the frequency point number of the acoustic calculation, is the weight coefficient of the frequency f, is the material sound absorption coefficient of the selected material at the frequency f extracted from the material selection result, is the sound absorption coefficient of the original material of the connected defect region at the frequency f, is the average sound wave incidence angle of the connected defect region, is the frequency correction factor, is the reference frequency.

8. A classroom acoustic environment optimization design apparatus, characterized by, The device comprises: a demand analysis and model acquisition module configured to perform semantic analysis processing on a natural language demand text input by a user, generate a set of quantized acoustic target values, and acquire a preset classroom geometric model; a sound field simulation and defect analysis module configured to perform sound field simulation processing through a hybrid acoustic model based on the set of quantized acoustic target values and the classroom geometric model, generate a three-dimensional acoustic parameter matrix, and perform point-by-point comparison processing on the three-dimensional acoustic parameter matrix and the set of quantized acoustic target values, calculate a reverberation time deviation value and a speech intelligibility deviation value, and generate an acoustic defect benchmark map; a material selection and layout generation module configured to perform spatial constraint clustering analysis processing on the acoustic defect benchmark map, generate a defect area feature table, select materials through a material matching rule library based on the defect area feature table, obtain a material selection result, and calculate a minimum laying area in combination with the defect area feature table to generate a material layout scheme; a scheme verification and output module configured to perform sound field re-simulation processing based on the material layout scheme, generate a new acoustic parameter matrix, perform comprehensive deviation value calculation processing on the new acoustic parameter matrix and the set of quantized acoustic target values to obtain a comprehensive deviation value, and output the corresponding material layout scheme as a final scheme when the comprehensive deviation value meets a preset threshold value. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to implement the steps of the method of any one of claims 1 to 7.

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

Images