Intelligent calibration method, device and equipment for pickup distance of microphone pickup and medium

Abstract

The invention relates to the technical field of electronic information, in particular to an intelligent calibration method and device for the pickup distance of a microphone pickup, equipment and a medium, and finally realizes automatic adjustment of microphone configuration through multi-sensor data acquisition and processing, construction of a dynamic acoustic model through a hybrid modeling algorithm, adaptive structure optimization and a nonlinear optimization strategy. Therefore, the pickup distance is optimized. According to the invention, the pickup precision and stability of the microphone are greatly improved, intelligent response can be realized in different environments, and high-quality audio acquisition and transmission are realized.

Description

Technical Field

[0001] The present invention relates to the field of electronic information technology, and in particular to a method, device, equipment and medium for intelligently calibrating the sound pickup distance of a microphone pickup. Background Art

[0002] With the widespread use of wireless microphones in teaching, corporate meetings and other occasions, the quality of sound pickup directly affects the audience's experience. Currently, wireless microphones on the market have low sensitivity and the distance between the microphone and the sound source is uncertain, resulting in significant differences in sound pickup effects. Existing technologies mainly determine the sound pickup distance through simple signal-to-noise ratio calculations and static distance measurements. This method is easily affected by environmental noise and changes in the user's position, resulting in unstable sound pickup, which in turn affects sound quality and communication efficiency.

[0003] The prior art (Chinese invention patent, publication number: CN114679647B, name: Method, device, equipment and readable storage medium for determining the pickup distance of a wireless microphone) mainly uses image recognition and signal-to-noise ratio calculation methods to estimate the distance between the microphone and the user's lips through the video or image data of the wireless microphone. This technology has significant deficiencies in noise processing and distance estimation. First, image recognition technology relies on clear image quality and stable lighting conditions, and changes in ambient light may lead to recognition errors. Secondly, the signal-to-noise ratio calculation only provides a static distance assessment and cannot dynamically adapt to user movement or environmental changes. In addition, the prior art lacks the ability to respond to real-time environmental changes, resulting in poor sound pickup effects. Summary of the invention

[0004] In view of the many problems existing in the above-mentioned prior art, the present invention provides a method, device, equipment and medium for intelligent calibration of the pickup distance of a microphone pickup. The present invention realizes dynamic intelligent calibration of the microphone pickup distance through multi-sensor fusion and intelligent algorithm optimization. First, infrared, acoustic and pressure sensors are used to obtain multi-dimensional environmental and user data. After denoising and standardization, a hybrid modeling algorithm is used to construct a dynamic acoustic model. Then, the microphone configuration is adjusted in real time through an adaptive structural optimization algorithm so that it can achieve the best pickup effect under various environmental conditions. Finally, combined with a multi-mode optimization algorithm, the robustness and adaptability of the system are improved to ensure the stability and efficiency of the pickup quality.

[0005] The method for intelligently calibrating the pickup distance of a microphone pickup comprises the following steps:

[0006] Use infrared sensors, array microphones and pressure sensors to collect data, including initial distance data, sound source direction data and holding status data, and perform denoising and standardization on the collected data to obtain comprehensive feature data;

[0007] A hybrid modeling algorithm based on physical and statistical methods is used to build a dynamic acoustic model based on comprehensive feature data to generate dynamic sound field data. The dynamic sound field data is analyzed through an adaptive structural optimization algorithm, and a calibration strategy is selected and executed to generate real-time calibration data.

[0008] Based on real-time calibration data and dynamic sound field data, and in accordance with specific optimization conditions, a nonlinear optimization algorithm combining genetic algorithm, simulated annealing and deep reinforcement learning is used for multi-mode optimization to generate the final optimization parameter data. The final optimization parameter data is analyzed using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration.

[0009] Preferably, the data collection step comprises the following steps:

[0010] Acquire initial distance data through an infrared sensor, and the initial distance data is used to detect the spatial distance between the user and the microphone;

[0011] Acquire sound source direction data through array microphones, which are used to identify the source direction of the sound and the sound field distribution;

[0012] The holding status data is obtained through the pressure sensor, and the holding status data is used to monitor the holding force and angle of the microphone to correct the impact of holding deviation on sound pickup.

[0013] Preferably, the generation of the comprehensive feature data comprises the following steps:

[0014] Applying a noise removal algorithm to the initial distance data, sound source direction data, and holding state data to remove environmental noise and signal interference;

[0015] The denoised data is standardized, and the normalization technology is used to make different data dimensions have consistent ranges and units, generating standardized initial distance data, standardized sound source direction data, and standardized holding status data;

[0016] The standardized data is comprehensively processed using feature fusion technology to form comprehensive feature data to comprehensively reflect the environmental acoustic characteristics and user interaction characteristics.

[0017] Preferably, the construction of the dynamic acoustic model comprises the following steps:

[0018] The basic acoustic model is constructed based on the physical acoustic characteristics, and the propagation characteristics of sound waves in space are simulated by the acoustic equation. The acoustic equation is:

[0019]

[0020] in, is the Laplace operator, Ψ(r, t) represents the sound pressure, r is the position vector, t is the time, and c is the speed of sound;

[0021] Use statistical analysis techniques to dynamically modify the basic acoustic model to generate a corrected acoustic model to adapt to environmental changes;

[0022] Combined with real-time environmental feedback data, the corrected acoustic model is further adjusted by iteratively updating the model parameters to generate dynamic sound field data to reflect the actual acoustic characteristics of the current environment.

[0023] Preferably, the adaptive structural optimization algorithm selects the calibration strategy including the following steps:

[0024] Perform sound field structure analysis on dynamic sound field data and identify acoustic patterns and change trends in the environment through pattern recognition technology;

[0025] Based on the identified acoustic patterns and changing trends, select a calibration strategy that adapts to the current environment, including: adjusting the microphone's pickup distance and directivity to generate real-time calibration data.

[0026] Preferably, the nonlinear optimization algorithm comprises the following steps:

[0027] The real-time calibration data is preliminarily optimized by genetic algorithm, the natural selection process is simulated to screen the optimal parameter combination, and the initial optimized parameter data is generated, where the fitness function f(x) is used to evaluate the quality of each individual x;

[0028] Use the simulated annealing algorithm to search for the global optimal solution among multiple possible parameter combinations, and avoid falling into local extreme values ​​by controlling the annealing process. The temperature adjustment function is:

[0029] T(t)=T0×α t

[0030] Where T(t) is the current temperature, T0 is the initial temperature, α is the cooling coefficient, and t is the number of iterations;

[0031] Deep reinforcement learning technology is used to deeply optimize the intermediate optimization parameter data. Through intelligent learning environment response behavior, the adaptability of the optimization parameters is continuously improved. The reward function of strategy θ is:

[0032] R(s,a)

[0033] Among them, R(s, a) is the reward value for taking action a in state s;

[0034] Generate final optimized parameter data.

[0035] Preferably, the generation of the comprehensive calibration decision data comprises the following steps:

[0036] The correlation between the final optimized parameter data and the dynamic sound field data is analyzed by hybrid logistic regression to generate logical calibration recommendation data, which is used to predict and evaluate the effectiveness of the calibration strategy;

[0037] Combine the logical calibration suggestion data with the dynamic sound field data, analyze and reason the uncertain factors through the fuzzy reasoning system, and generate fuzzy calibration suggestion data, which is used to deal with the uncertainty and changes in the dynamic environment;

[0038] Combining logical calibration suggestion data and fuzzy calibration suggestion data, comprehensive calibration decision data is generated through interactive calculation to optimize microphone configuration and realize pickup distance calibration.

[0039] A device for intelligently calibrating the sound pickup distance of a microphone pickup, comprising:

[0040] A data acquisition module is used to collect data using infrared sensors, array microphones and pressure sensors to generate initial distance data, sound source direction data and holding state data, and to perform denoising and standardization on the collected data to obtain comprehensive feature data;

[0041] The dynamic modeling module is used to construct a dynamic acoustic model and generate dynamic sound field data based on comprehensive feature data using a hybrid modeling algorithm based on a combination of physical and statistical methods;

[0042] A calibration strategy module, used to analyze dynamic sound field data through an adaptive structural optimization algorithm, select and execute a calibration strategy to generate real-time calibration data;

[0043] The optimization module is used to perform multi-mode optimization based on real-time calibration data and dynamic sound field data and specific optimization conditions by applying a nonlinear optimization algorithm combining genetic algorithm, simulated annealing and deep reinforcement learning to generate final optimization parameter data;

[0044] The decision module is used to analyze the final optimized parameter data using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration.

[0045] An electronic device, comprising:

[0046] at least one processor;

[0047] The memory stores computer-readable instructions, and when the instructions are executed by the processor, the processor executes the intelligent calibration method for the microphone pickup distance.

[0048] A computer-readable storage medium stores computer-readable instructions, which, when executed by a computing device, enable the computing device to execute any of the microphone pickup distance intelligent calibration methods.

[0049] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0050] The present invention realizes precise and intelligent calibration of the sound pickup distance by combining multiple sensor technologies, and uses infrared sensors, array microphones and pressure sensors to collect initial distance data, sound source direction data and holding status data to form comprehensive feature data;

[0051] The present invention constructs a dynamic acoustic model through a hybrid modeling algorithm based on a combination of physical and statistical methods to generate real-time dynamic sound field data;

[0052] The present invention applies nonlinear optimization algorithms such as genetic algorithms, simulated annealing and deep reinforcement learning to generate final optimization parameters and realize automatic adjustment of microphone configuration, thereby optimizing the pickup distance;

[0053] The process of the present invention not only effectively solves the defects of the prior art in noise processing and real-time response, but also greatly improves the accuracy and stability of sound pickup. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of the process of the present invention;

[0055] Figure 2 It is a schematic diagram of the sensor data fusion process in the present invention;

[0056] Figure 3 This is a flow chart of the optimization algorithm in the present invention;

[0057] Figure 4 It is a structural block diagram of the device of the present invention. DETAILED DESCRIPTION

[0058] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0059] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0060] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0061] like Figure 1 As shown, the method for intelligently calibrating the pickup distance of a microphone pickup comprises the following steps:

[0062] like Figure 2 As shown, infrared sensors, array microphones and pressure sensors are used to collect data, including initial distance data, sound source direction data and holding state data, and the collected data is denoised and standardized to obtain comprehensive feature data;

[0063] Preferably, the data collection step comprises the following steps:

[0064] Acquire initial distance data through an infrared sensor, and the initial distance data is used to detect the spatial distance between the user and the microphone;

[0065] Acquire sound source direction data through array microphones, which are used to identify the source direction of the sound and the sound field distribution;

[0066] The holding status data is obtained through the pressure sensor, and the holding status data is used to monitor the holding force and angle of the microphone to correct the impact of holding deviation on sound pickup.

[0067] In the present invention, the data collection step ensures the accuracy and real-time performance of the sound pickup distance calibration by combining multiple sensors. This step mainly includes the coordinated work of infrared sensors, array microphones and pressure sensors. These devices collect different types of data and fuse them through algorithms to provide a solid data foundation for subsequent intelligent calibration.

[0068] By obtaining initial distance data through the infrared sensor, the infrared sensor can accurately measure the spatial distance between the user and the microphone. This is because the infrared sensor determines the distance by emitting infrared light waves and receiving reflected signals, calculating the time it takes for the light waves to propagate in the air. This measurement method has high accuracy and reliability within a short distance and can provide dynamically changing distance information in real time. For the intelligent calibration of the microphone, accurate distance data is the key because it directly affects the time delay and attenuation of the sound wave propagating to the microphone, thereby affecting the quality of sound pickup.

[0069] Acquire sound source direction data through array microphone. The array microphone is composed of multiple microphone units, and the direction of the sound source and the sound field distribution can be determined by a differential algorithm. The principle is to use the time difference of the sound wave reaching each microphone unit, and the position and direction of the sound source can be determined by calculating these time differences. The application of this technology in the present invention is crucial, because the sound source direction data is not only used to adjust the directivity of the microphone to optimize the sound pickup effect, but also used to analyze the distribution characteristics of the ambient sound field, thereby providing important input for the dynamic acoustic model. The application of this multi-unit microphone can effectively filter background noise and enhance the clarity and resolution of the target sound signal.

[0070] The holding status data is obtained through the pressure sensor. The pressure sensor is used to detect the force and angle when the microphone is held. This data helps to identify and correct the pickup deviation caused by improper holding. The pressure sensor measures the contact force distribution and angle changes of the holding part, and feeds back to the system for real-time adjustment. This is particularly important for the user experience of the pickup, especially in mobile scenes or dynamic environments. Changes in the holding state may cause the microphone position to shift, thereby affecting the pickup quality. Through real-time monitoring and adjustment, sound distortion or deviation caused by unstable holding can be effectively reduced.

[0071] Preferably, the generation of the comprehensive feature data comprises the following steps:

[0072] Applying a noise removal algorithm to the initial distance data, sound source direction data, and holding state data to remove environmental noise and signal interference;

[0073] The denoised data is standardized, and the normalization technology is used to make different data dimensions have consistent ranges and units, generating standardized initial distance data, standardized sound source direction data, and standardized holding status data;

[0074] The standardized data is comprehensively processed using feature fusion technology to form comprehensive feature data to comprehensively reflect the environmental acoustic characteristics and user interaction characteristics.

[0075] Apply a noise removal algorithm to the initial distance data, sound source direction data, and grip state data. This step aims to eliminate environmental noise and signal interference to ensure data purity and accuracy. The noise removal algorithm usually uses filtering techniques, such as Kalman filtering or wavelet transform, which can effectively identify and filter out unnecessary signal components. This is especially important for the pickup, because noise interference may lead to incorrect distance measurement and direction judgment, thus affecting the accuracy of calibration. Through denoising, the system can obtain more reliable basic data, thereby improving the effectiveness of subsequent processing steps.

[0076] Standardize the denoised data. Standardization is an indispensable part of data preprocessing. Through normalization technology, different data dimensions have consistent ranges and units. Specifically, standardization can use linear scaling or Z-score methods to make each data type within the same order of magnitude to avoid deviations caused by differences in data scales. For the calibration process of the microphone, the standardized initial distance data, sound source direction data, and holding status data can be uniformly processed and analyzed to ensure that the mutual influence between different data sources is reasonably evaluated. This processing can not only reduce data deviations, but also improve the efficiency and accuracy of data analysis.

[0077] The standardized data is processed comprehensively using feature fusion technology to form comprehensive feature data. Feature fusion is to generate a more representative feature set by integrating information from multiple data sources. In the present invention, feature fusion technology may involve methods such as weighted average, principal component analysis or neural network fusion. Through fusion, the comprehensive feature data can fully reflect the environmental acoustic characteristics and the user's interactive characteristics, and provide multi-dimensional information support for the intelligent calibration algorithm of the microphone. This fusion not only improves the descriptive ability of the data, but also enhances the system's perception and adaptability to environmental changes.

[0078] For example, in a complex audio environment, the initial distance data may indicate that the user is far from the microphone, while the sound source direction data indicates that the sound comes from multiple directions. Through feature fusion, the system can recognize that in this case it may be necessary to increase the sensitivity of the microphone and adjust the pickup direction to ensure that all important sound signals can be accurately captured. In this way, the generation of comprehensive feature data directly provides a basis for subsequent calibration strategies, enabling the entire pickup system to achieve efficient and accurate sound collection in a dynamic environment.

[0079] A hybrid modeling algorithm based on physical and statistical methods is used to build a dynamic acoustic model based on comprehensive feature data to generate dynamic sound field data. The dynamic sound field data is analyzed through an adaptive structural optimization algorithm, and a calibration strategy is selected and executed to generate real-time calibration data.

[0080] The goal of this step is to optimize the microphone configuration to improve the sound pickup accuracy through a deep understanding of the environmental acoustic characteristics and real-time response.

[0081] A dynamic acoustic model is constructed using comprehensive feature data. The comprehensive feature data contains information from multiple sources, such as distance, sound source direction, and holding state. These data are processed in advance to form a multi-dimensional feature vector. On this basis, the hybrid modeling algorithm combines the advantages of physical models and statistical models. Physical models are usually based on the basic principles of sound wave propagation, taking into account the physical properties of sound when it propagates in the air, such as attenuation, reflection, and refraction, and can provide an accurate description of the ambient sound field. However, due to the complexity of the actual environment, physical models alone may not be able to fully capture all acoustic features. Therefore, the statistical model uses a data-driven approach to compensate and correct acoustic changes in the actual environment, and can adapt to different environmental conditions and dynamic changes.

[0082] Through the hybrid modeling algorithm, the system is able to generate dynamic sound field data that not only reflects the acoustic state of the current environment, but also includes predictions of future changes in the sound field. This dynamic sound field data provides key basic information for intelligent calibration. For example, in an open space, the acoustic model can identify the main sound reflection paths and attenuation patterns, thereby helping the pickup optimize its sensitivity and directivity.

[0083] The adaptive structural optimization algorithm analyzes the dynamic sound field data and selects and executes the calibration strategy. The core of the adaptive structural optimization algorithm lies in its flexibility and responsiveness. It can analyze the changes in the sound field data in real time and dynamically adjust the calibration strategy. Specifically, the algorithm first analyzes the dynamic sound field data to identify the key acoustic features and changing trends in the environment. Based on these analysis results, the system can intelligently select the calibration strategy that best suits the current sound field state. These strategies may include adjusting the microphone's directivity, gain, and pickup pattern.

[0084] Once the calibration strategies are determined, the system will immediately execute them and generate real-time calibration data. This data is not only used to adjust the current pickup configuration, but also to monitor the calibration effect in real time to ensure that the system always maintains the best pickup performance in different environments. Through this adaptive calibration, the system can effectively respond to environmental changes and ensure efficient and accurate sound collection in various complex acoustic environments.

[0085] For example, in a conference room environment, the hybrid modeling algorithm can identify the impact of walls and furniture on sound propagation, and adjust the microphone's pickup direction and sensitivity through an adaptive optimization strategy to ensure that the voices of all participants can be clearly captured. Through this dynamic response mechanism, the system can adapt to environmental changes in real time, improving the pickup quality and user experience.

[0086] Preferably, the construction of the dynamic acoustic model comprises the following steps:

[0087] The basic acoustic model is constructed based on the physical acoustic characteristics, and the propagation characteristics of sound waves in space are simulated by the acoustic equation. The acoustic equation is:

[0088]

[0089] in, is the Laplace operator, Ψ(r, t) represents the sound pressure, r is the position vector, t is the time, and c is the speed of sound;

[0090] Use statistical analysis techniques to dynamically modify the basic acoustic model to generate a corrected acoustic model to adapt to environmental changes;

[0091] Combined with real-time environmental feedback data, the corrected acoustic model is further adjusted by iteratively updating the model parameters to generate dynamic sound field data to reflect the actual acoustic characteristics of the current environment.

[0092] In the present invention, the construction of a dynamic acoustic model is one of the important steps to achieve accurate calibration. This process combines physical acoustic characteristics and statistical analysis techniques to establish an acoustic model that can dynamically adapt to environmental changes, thereby improving the microphone's sound pickup accuracy and adaptability.

[0093] This equation describes the wave characteristics of sound waves when they propagate in space. The curvature of the sound pressure field is calculated and combined with the time derivative to describe the change of sound pressure over time. This model can simulate the propagation path and characteristics of sound waves in different media, providing a basis for understanding the acoustic characteristics of the environment.

[0094] The acoustic characteristics in the actual environment are often affected by many factors, such as temperature, humidity, obstacles, etc., which can cause changes in the propagation path and speed of sound waves. Therefore, the physical acoustic model alone may not be able to fully describe the actual acoustic characteristics in the environment. To this end, the present invention introduces statistical analysis technology to dynamically correct the basic acoustic model. By analyzing historical data and real-time feedback, the impact of environmental changes on the sound field is identified, and the model parameters are adjusted to generate a corrected acoustic model. This process enables the model to more accurately reflect the current acoustic environment, thereby improving the accuracy of calibration.

[0095] Combined with real-time environmental feedback data, the present invention iteratively updates model parameters, further adjusts the corrected acoustic model, and generates dynamic sound field data. These dynamic sound field data not only reflect the current acoustic state, but also can predict the changing trend of environmental acoustic characteristics. For example, in a conference room, as people move and their positions change, the propagation path of sound waves will also change. By generating dynamic sound field data, the system can adjust the directivity and sensitivity of the microphone in real time to adapt to the new sound field environment.

[0096] Through this dynamic adjustment mechanism, the system can respond to environmental changes in real time, ensuring that the microphone always maintains the best sound pickup performance under different acoustic conditions. Ultimately, this dynamic acoustic modeling method that combines physical models and statistical analysis not only improves the adaptability and accuracy of the microphone, but also enhances the system's understanding and processing capabilities of complex acoustic environments, achieving more efficient and accurate sound collection results.

[0097] Preferably, the adaptive structural optimization algorithm selects the calibration strategy including the following steps:

[0098] Perform sound field structure analysis on dynamic sound field data and identify acoustic patterns and change trends in the environment through pattern recognition technology;

[0099] Based on the identified acoustic patterns and changing trends, select a calibration strategy that adapts to the current environment, including: adjusting the microphone's pickup distance and directivity to generate real-time calibration data.

[0100] In the present invention, an adaptive structure optimization algorithm is used to select and execute a calibration strategy, which is a key step in achieving efficient sound pickup in a dynamic environment. This step analyzes dynamic sound field data, identifies environmental acoustic patterns, and adjusts the configuration of the microphone in real time to optimize the sound pickup effect.

[0101] The present invention uses dynamic sound field data to perform sound field structure analysis. Dynamic sound field data contains acoustic information in the real-time environment, such as the propagation path of sound waves, intensity distribution, and sound source location. Through pattern recognition technology, the system can identify acoustic patterns and change trends in the environment. Pattern recognition technology usually includes steps such as feature extraction, classification, and clustering, which are used to analyze structural features in sound field data. This analysis can reveal the reflection, absorption, and scattering characteristics in the environment, thereby helping the system understand the complexity of the current acoustic environment.

[0102] For example, in a theater environment, the system can identify the main sound reflection paths and the impact of the audience's movement on the sound field through sound field structure analysis. By identifying these features, the system can predict possible acoustic change trends, such as sound field fluctuations caused by audience entry or exit. This analysis provides an important basis for the selection of calibration strategies.

[0103] Based on the identified acoustic patterns and changing trends, the system selects calibration strategies that adapt to the current environment. These strategies may include adjusting the microphone's pickup distance, directivity, and gain to generate real-time calibration data. Adjusting the pickup distance ensures that the microphone can accurately pick up sounds from different distances, while adjusting the directivity optimizes the microphone's receiving angle, enabling it to focus on the target sound source, thereby reducing interference from background noise.

[0104] Through the adaptive structural optimization algorithm, the system is able to execute the selected calibration strategy in real time and dynamically adjust according to environmental changes. For example, in an open office space, the sound field characteristics may change frequently as the flow of people and activities change. The adaptive structural optimization algorithm can identify these changes and adjust the configuration of the microphone in time to ensure the stability and high quality of the pickup effect.

[0105] This adaptive calibration mechanism enables the microphone to always maintain the best sound pickup performance in a changing environment through real-time analysis and adjustment. Finally, the present invention realizes efficient sound pickup calibration in a dynamic environment by combining sound field structure analysis and adaptive optimization strategy, which significantly improves the system's adaptability to complex acoustic environments and sound pickup quality.

[0106] like Figure 3 As shown, based on the real-time calibration data and dynamic sound field data, according to specific optimization conditions, a nonlinear optimization algorithm combining genetic algorithm, simulated annealing and deep reinforcement learning is used to perform multi-mode optimization to generate final optimization parameter data; the final optimization parameter data is analyzed using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration.

[0107] In the present invention, the application of nonlinear optimization algorithm is an important step to ensure that the system can achieve the best sound pickup performance in a dynamic environment. The process combines genetic algorithm, simulated annealing and deep reinforcement learning to perform multi-mode optimization and finally generate the optimal sound pickup configuration parameters.

[0108] Based on real-time calibration data and dynamic sound field data, the system identifies the acoustic features in the current environment and determines the specific goals and constraints for optimization. These data provide a detailed description of the current acoustic environment, including factors such as the location of the sound source, the distribution of the sound field, and the ambient noise. The optimization process is based on these data and applies a variety of optimization techniques such as genetic algorithms, simulated annealing, and deep reinforcement learning to find the best combination of calibration parameters.

[0109] Genetic algorithms simulate biological evolution and perform a global search of parameter space through operations such as selection, crossover, and mutation. The advantage of genetic algorithms lies in their powerful global search capabilities, which can find multiple potential optimal solutions in a huge parameter space. Subsequently, simulated annealing algorithms further refine these solutions, simulating the physical annealing process and gradually reducing the "temperature" to converge to the global optimal solution. Simulated annealing controls the cooling process to avoid falling into local extremes and ensure that the global optimal solution is found. Finally, deep reinforcement learning intelligently adjusts the effects of different calibration parameter combinations by learning environmental responses. Deep reinforcement learning uses a reward mechanism to continuously optimize parameter combinations, thereby maintaining the flexibility and adaptability of the pickup configuration in a dynamically changing environment.

[0110] After the multi-mode optimization generates the final optimized parameter data, the present invention uses a hybrid logistic regression and fuzzy reasoning system to conduct a comprehensive analysis of these data. Hybrid logistic regression is used to evaluate the mutual influence and correlation between parameters, and provide basic logical analysis for calibration decisions. Through logistic regression, the system can predict the potential impact of different parameter combinations on the pickup effect. The fuzzy reasoning system handles uncertainty, analyzes and infers uncertain factors in complex environments through fuzzy sets and rule systems, and generates comprehensive calibration decision data. This fuzzy reasoning can effectively deal with noise and nonlinear characteristics in complex environments, ensuring that the system is highly robust to environmental changes.

[0111] Finally, the comprehensive calibration decision data is used to adjust the configuration of the microphone to achieve the final pickup distance calibration. Through these steps, the system can dynamically adjust the configuration parameters of the microphone in a changing acoustic environment to ensure the optimization and stability of the pickup performance. For example, in a noisy indoor environment, the optimization algorithm can adjust the pickup direction and sensitivity of the microphone to enhance the clarity of the target sound while reducing the interference of background noise. This dynamic optimization process significantly improves the adaptability and pickup accuracy of the system, ensuring efficient and accurate sound collection under various environmental conditions. Through this comprehensive optimization, the present invention successfully achieves the efficiency and reliability of intelligent calibration of the pickup distance.

[0112] Preferably, the nonlinear optimization algorithm comprises the following steps:

[0113] The real-time calibration data is preliminarily optimized by genetic algorithm, the natural selection process is simulated to screen the optimal parameter combination, and the initial optimized parameter data is generated, where the fitness function f(x) is used to evaluate the quality of each individual x;

[0114] Use the simulated annealing algorithm to search for the global optimal solution among multiple possible parameter combinations, and avoid falling into local extreme values ​​by controlling the annealing process. The temperature adjustment function is:

[0115] T(t)=T0×αt

[0116] Where T(t) is the current temperature, T0 is the initial temperature, α is the cooling coefficient, and t is the number of iterations;

[0117] Deep reinforcement learning technology is used to deeply optimize the intermediate optimization parameter data. Through intelligent learning environment response behavior, the adaptability of the optimization parameters is continuously improved. The reward function of strategy θ is:

[0118] R(s,a)

[0119] Among them, R(s, a) is the reward value for taking action a in state s;

[0120] Generate final optimized parameter data.

[0121] In the present invention, the nonlinear optimization algorithm aims to achieve comprehensive optimization of microphone configuration parameters through the combination of genetic algorithm, simulated annealing and deep reinforcement learning, ensuring the best sound pickup effect under various environmental conditions.

[0122] First, the real-time calibration data is preliminarily optimized using a genetic algorithm. A genetic algorithm is a global optimization technique that simulates the natural selection process and selects the optimal parameter combination through selection, crossover, and mutation operations. In this process, each individual x represents a parameter combination, and its quality is evaluated by the fitness function f(x). The design of the fitness function is based on a comprehensive evaluation of the sound pickup effect, such as signal-to-noise ratio, sensitivity, and response speed. Through multiple generations of iterations, the genetic algorithm can effectively explore the parameter space and identify potential excellent parameter combinations. This preliminary optimization lays the foundation for subsequent refined optimization.

[0123] After obtaining the initial optimization parameter data, the present invention applies the simulated annealing algorithm for further optimization. Simulated annealing searches for the global optimal solution among multiple possible parameter combinations by simulating the physical annealing process. The algorithm reduces the search range by controlling the gradual reduction of the "temperature" to avoid falling into local extreme values. As in the above temperature adjustment function, T(t) represents the current temperature, T0 is the initial temperature, α is the cooling coefficient, and t is the number of iterations. Through this cooling mechanism, simulated annealing can refine the parameter selection to ensure that the optimal pickup configuration is found.

[0124] After generating the intermediate optimization parameter data, the present invention further applies deep reinforcement learning technology for deep optimization. Deep reinforcement learning is an intelligent algorithm that optimizes decision-making by learning environmental feedback. In this process, the reward function R(s, a) of strategy θ is used to evaluate the effect of taking action a in state s. The reward value is evaluated based on the impact of the action on the sound pickup effect, such as whether the sound clarity is improved or the background noise is reduced. By continuously adjusting the strategy, deep reinforcement learning can dynamically optimize the parameter settings so that the system maintains efficient sound pickup performance in a changing acoustic environment.

[0125] Through the collaborative work of these three algorithms, the optimized parameter data generated is used to adjust the microphone configuration in real time. In a complex audio environment, such as an open office space, the microphone may need to dynamically adjust its pickup direction and sensitivity to adapt to changes in background noise and crowd movement. This nonlinear optimization process ensures that the microphone can provide high-quality sound collection in a variety of environments.

[0126] Preferably, the generation of the comprehensive calibration decision data comprises the following steps:

[0127] The correlation between the final optimized parameter data and the dynamic sound field data is analyzed by hybrid logistic regression to generate logical calibration recommendation data, which is used to predict and evaluate the effectiveness of the calibration strategy;

[0128] Combine the logical calibration suggestion data with the dynamic sound field data, analyze and reason the uncertain factors through the fuzzy reasoning system, and generate fuzzy calibration suggestion data, which is used to deal with the uncertainty and changes in the dynamic environment;

[0129] Combining logical calibration suggestion data and fuzzy calibration suggestion data, comprehensive calibration decision data is generated through interactive calculation to optimize microphone configuration and realize pickup distance calibration.

[0130] In the present invention, the generation of comprehensive calibration decision data is a key step to ensure that the system can dynamically respond to environmental changes and achieve accurate calibration. The process analyzes and optimizes the configuration of the microphone to achieve the best sound pickup effect through the combination of hybrid logistic regression and fuzzy inference system.

[0131] The system uses hybrid logistic regression to analyze the correlation between the final optimized parameter data and the dynamic sound field data. Hybrid logistic regression is a statistical analysis technique that can reveal linear and nonlinear relationships between variables. In this step, logistic regression is used to evaluate the impact of parameter adjustments on sound field characteristics and generate logical calibration recommendation data. This data reflects the potential impact of different parameter combinations on pickup performance and is the basis for predicting and evaluating calibration strategies. For example, by analyzing the relationship between sound field data and optimized parameters, logistic regression can identify which parameter adjustments can significantly improve the signal-to-noise ratio or enhance the clarity of the target sound.

[0132] The system combines the logical calibration suggestion data with the dynamic sound field data, and analyzes and infers the uncertain factors through the fuzzy reasoning system. The fuzzy reasoning system is a computational method based on fuzzy set theory, which is good at processing fuzzy and uncertain information. In a complex acoustic environment, noise and environmental changes are often uncertain. The fuzzy reasoning system analyzes these uncertain factors through fuzzy rules and reasoning mechanisms to generate fuzzy calibration suggestion data. For example, in an open environment, the changes in background noise may be difficult to accurately model. Fuzzy reasoning can provide reasonable calibration suggestions under such uncertainty to cope with dynamic changes.

[0133] The system combines the logical calibration suggestion data and the fuzzy calibration suggestion data to generate comprehensive calibration decision data through interactive calculation. This step generates a comprehensive calibration decision by integrating the results of logical analysis and fuzzy reasoning. The decision is used to adjust the microphone configuration in real time to achieve accurate pickup distance calibration. Through this interactive calculation, the system can comprehensively consider the influence of various factors and formulate the optimal pickup configuration strategy.

[0134] For example, in a multi-functional conference room, the microphone may need to be constantly adjusted to accommodate different sound sources and activity types. By integrating calibration decision data, the system can dynamically adjust the microphone's pickup direction, sensitivity and gain to ensure efficient sound capture in a variety of situations.

[0135] like Figure 4 As shown, a device for intelligently calibrating the sound pickup distance of a microphone pickup comprises:

[0136] The data acquisition module is used to collect data using infrared sensors, array microphones and pressure sensors, generate initial distance data, sound source direction data and holding state data, and perform denoising and standardization on the collected data to obtain comprehensive feature data; the data acquisition module is the basic part of the present invention, and is responsible for obtaining multi-dimensional data of the environment and the user through infrared sensors, array microphones and pressure sensors. The infrared sensor is used to measure the initial distance data and provide spatial distance information between the user and the microphone; the array microphone is responsible for obtaining the sound source direction data, identifying the source direction of the sound and the sound field distribution; the pressure sensor is used to detect the holding state data and monitor the holding force and angle of the microphone. Through these sensors, the system can obtain comprehensive initial environmental data. Subsequently, the data acquisition module performs denoising and standardization on these data to eliminate environmental noise and signal deviation to ensure the accuracy and consistency of the data. The processed comprehensive feature data provides a reliable basis for subsequent dynamic modeling and optimization.

[0137] The dynamic modeling module is used to use a hybrid modeling algorithm based on a combination of physical and statistical methods to build a dynamic acoustic model based on comprehensive feature data and generate dynamic sound field data; the dynamic modeling module uses comprehensive feature data to build a dynamic acoustic model through a hybrid modeling algorithm based on a combination of physical and statistical methods. This process combines physical acoustic characteristics and statistical analysis techniques to accurately simulate the propagation characteristics of sound waves in the current environment. The physical model takes into account the reflection, refraction and attenuation characteristics of sound waves, while the statistical model is dynamically corrected based on actual data to adapt to environmental changes. In this way, the system generates dynamic sound field data that can accurately reflect the acoustic state in the current environment. The dynamic sound field data provides an important reference for subsequent calibration strategies.

[0138] The calibration strategy module is used to analyze the dynamic sound field data through the adaptive structure optimization algorithm, select and execute the calibration strategy to generate real-time calibration data; the calibration strategy module analyzes the dynamic sound field data through the adaptive structure optimization algorithm and selects the calibration strategy that adapts to the current environment. The adaptive structure optimization algorithm is flexible and responsive, and can identify key features in the sound field in real time and select the best calibration scheme. This includes adjusting the microphone's pickup distance, directivity, and sensitivity to generate real-time calibration data. Through real-time calibration, the system can quickly adapt to environmental changes to ensure the best pickup effect under dynamic conditions.

[0139] The optimization module is used to perform multi-mode optimization based on real-time calibration data and dynamic sound field data, according to specific optimization conditions, and apply a nonlinear optimization algorithm that combines genetic algorithms, simulated annealing, and deep reinforcement learning to generate the final optimized parameter data; the optimization module is based on real-time calibration data and dynamic sound field data, and applies a nonlinear optimization algorithm to perform multi-mode optimization. This module combines advanced optimization technologies such as genetic algorithms, simulated annealing, and deep reinforcement learning. Genetic algorithms explore the global optimal solution of parameter space by simulating natural selection; simulated annealing avoids falling into local extreme values ​​by controlling the annealing process; deep reinforcement learning uses environmental feedback to continuously adjust strategies and improve parameter adaptability. This multi-mode optimization process generates the final optimized parameter data, which provides a basis for further adjustment of microphone configuration.

[0140] The decision module is used to analyze the final optimized parameter data using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration. The decision module analyzes the final optimized parameter data through a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data. Hybrid logistic regression evaluates the correlation between parameters and predicts the pickup effect of different combinations; the fuzzy inference system handles the uncertainty in the environment and provides flexible decision support. Through interactive calculations, the system generates the final decision to optimize the microphone configuration and achieves accurate calibration of the pickup distance. In practical applications, this decision-making mechanism can dynamically adjust the configuration of the microphone to cope with complex and changing acoustic environments, ensuring that high-quality sound collection is always provided.

[0141] An electronic device, comprising:

[0142] at least one processor;

[0143] The memory stores computer-readable instructions, and when the instructions are executed by the processor, the processor executes the intelligent calibration method for the microphone pickup distance.

[0144] A computer-readable storage medium stores computer-readable instructions, which, when executed by a computing device, enable the computing device to execute any of the microphone pickup distance intelligent calibration methods.

[0145] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0146] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. An intelligent calibration method for microphone pickup distance, characterized in that: The following steps are involved: Use infrared sensors, array microphones and pressure sensors to collect data, including initial distance data, sound source direction data and holding status data, and perform denoising and standardization on the collected data to obtain comprehensive feature data; A hybrid modeling algorithm based on physical and statistical methods is used to build a dynamic acoustic model based on comprehensive feature data to generate dynamic sound field data. The dynamic sound field data is analyzed through an adaptive structural optimization algorithm, and a calibration strategy is selected and executed to generate real-time calibration data. Based on real-time calibration data and dynamic sound field data, according to specific optimization conditions, a nonlinear optimization algorithm combining genetic algorithm, simulated annealing and deep reinforcement learning is applied to perform multi-mode optimization to generate the final optimization parameter data; The final optimized parameter data is analyzed using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration.

2. The method for intelligently calibrating the pickup distance of a microphone according to claim 1, characterized in that: The data collection step comprises the following steps: Acquire initial distance data through an infrared sensor, and the initial distance data is used to detect the spatial distance between the user and the microphone; Acquire sound source direction data through array microphones, which are used to identify the source direction of the sound and the sound field distribution; The holding status data is obtained through the pressure sensor, and the holding status data is used to monitor the holding force and angle of the microphone to correct the impact of holding deviation on sound pickup.

3. The method for intelligently calibrating the pickup distance of a microphone according to claim 1, characterized in that: The generation of the comprehensive feature data comprises the following steps: Applying a noise removal algorithm to the initial distance data, sound source direction data, and holding state data to remove environmental noise and signal interference; The denoised data is standardized, and the normalization technology is used to make different data dimensions have consistent ranges and units, generating standardized initial distance data, standardized sound source direction data, and standardized holding status data; The standardized data is comprehensively processed using feature fusion technology to form comprehensive feature data to comprehensively reflect the environmental acoustic characteristics and user interaction characteristics.

4. The method for intelligently calibrating the pickup distance of a microphone according to claim 1, characterized in that: The construction of the dynamic acoustic model includes the following steps: The basic acoustic model is constructed based on the physical acoustic characteristics, and the propagation characteristics of sound waves in space are simulated by the acoustic equation. The acoustic equation is: in, is the Laplace operator, Ψ(r, t) represents the sound pressure, r is the position vector, t is the time, and c is the speed of sound; Use statistical analysis techniques to dynamically modify the basic acoustic model to generate a corrected acoustic model to adapt to environmental changes; Combined with real-time environmental feedback data, the corrected acoustic model is further adjusted by iteratively updating the model parameters to generate dynamic sound field data to reflect the actual acoustic characteristics of the current environment.

5. The method for intelligently calibrating the pickup distance of a microphone according to claim 1, characterized in that: The adaptive structural optimization algorithm selection calibration strategy comprises the following steps: Perform sound field structure analysis on dynamic sound field data and identify acoustic patterns and change trends in the environment through pattern recognition technology; Based on the identified acoustic patterns and changing trends, select a calibration strategy that adapts to the current environment, including: adjusting the microphone's pickup distance and directivity to generate real-time calibration data.

6. The method for intelligently calibrating the pickup distance of a microphone according to claim 1, characterized in that: The nonlinear optimization algorithm comprises the following steps: The real-time calibration data is preliminarily optimized by genetic algorithm, the natural selection process is simulated to screen the optimal parameter combination, and the initial optimized parameter data is generated, where the fitness function f(x) is used to evaluate the quality of each individual x; Use the simulated annealing algorithm to search for the global optimal solution among multiple possible parameter combinations, and avoid falling into local extreme values ​​by controlling the annealing process. The temperature adjustment function is: T(t)=T0×α t Where T(t) is the current temperature, T0 is the initial temperature, α is the cooling coefficient, and t is the number of iterations; Deep reinforcement learning technology is used to deeply optimize the intermediate optimization parameter data. Through intelligent learning environment response behavior, the adaptability of the optimization parameters is continuously improved. The reward function of strategy θ is: R(s,a) Among them, R(s, a) is the reward value for taking action a in state s; Generate final optimized parameter data.

7. The method for intelligently calibrating the pickup distance of a microphone according to claim 6, characterized in that: The generation of the comprehensive calibration decision data comprises the following steps: The correlation between the final optimized parameter data and the dynamic sound field data is analyzed by hybrid logistic regression to generate logical calibration recommendation data, which is used to predict and evaluate the effectiveness of the calibration strategy; Combine the logical calibration suggestion data with the dynamic sound field data, analyze and reason the uncertain factors through the fuzzy reasoning system, and generate fuzzy calibration suggestion data, which is used to deal with the uncertainty and changes in the dynamic environment; Combining logical calibration suggestion data and fuzzy calibration suggestion data, comprehensive calibration decision data is generated through interactive calculation to optimize microphone configuration and realize pickup distance calibration.

8. A device for intelligently calibrating the sound pickup distance of a microphone pickup, characterized in that: include: A data acquisition module is used to collect data using infrared sensors, array microphones and pressure sensors to generate initial distance data, sound source direction data and holding state data, and to perform denoising and standardization on the collected data to obtain comprehensive feature data; The dynamic modeling module is used to construct a dynamic acoustic model and generate dynamic sound field data based on comprehensive feature data using a hybrid modeling algorithm based on a combination of physical and statistical methods; A calibration strategy module, used to analyze dynamic sound field data through an adaptive structural optimization algorithm, select and execute a calibration strategy to generate real-time calibration data; The optimization module is used to perform multi-mode optimization based on real-time calibration data and dynamic sound field data and specific optimization conditions by applying a nonlinear optimization algorithm combining genetic algorithm, simulated annealing and deep reinforcement learning to generate final optimization parameter data; The decision module is used to analyze the final optimized parameter data using a hybrid logistic regression and fuzzy inference system to generate comprehensive calibration decision data to adjust the microphone configuration and achieve the final pickup distance calibration.

9. An electronic device, characterized in that: include: at least one processor; The memory stores computer-readable instructions, and when the instructions are executed by the processor, the processor executes the intelligent calibration method for the microphone pickup distance according to any one of claims 1-7.

10. A computer-readable storage medium having computer-readable instructions stored thereon, which, when executed by a computing device, causes the computing device to execute the method for intelligently calibrating the pickup distance of a microphone pickup as claimed in any one of claims 1 to 7.

Images