A construction site noise intelligent monitoring method and system

Abstract

The present application relates to the technical field of construction environment monitoring, in particular to a construction site noise intelligent monitoring method and system; the method comprises the following steps: taking the construction state constraint information as the prior condition, decomposing the noise signal into the construction machinery noise component and the background noise component; dynamically dividing the monitoring period into the construction period and the non-construction period, taking the noise signal of the non-construction period as the background noise reference value, calculating the corrected construction noise equivalent sound level value according to the difference between the total sound level of the construction period and the background noise sound level; calculating the noise contribution degree weight of each construction machinery, and outputting the comprehensive noise emission index; the system comprises the following modules: a multi-source sensing data alignment module, a noise signal feature source separation module, a background noise correction module, and a noise contribution degree evaluation module; through the above-mentioned mode, the effective construction noise and the non-source background noise are distinguished, the monitoring precision and the intelligent level are improved, and thus the existing construction site environmental protection monitoring standard is met.

Description

Technical Field

[0001] This invention relates to the field of construction environment monitoring technology, and in particular to an intelligent method and system for monitoring construction site noise. Background Technology

[0002] With the rapid development of the urban infrastructure industry, the number of construction projects, municipal renovations, and rail transit construction projects continues to increase. Construction site noise is characterized by its suddenness, randomness, high intensity, and multiple sound sources superimposed on each other. Therefore, all-weather, high-precision, and intelligent monitoring of construction site noise is a core requirement for construction environmental protection management.

[0003] Currently, most mainstream construction site noise monitoring technologies use a single acoustic sensor to collect noise data and perform noise monitoring and exceedance determination through simple sound level calculations. This technical solution has several inherent drawbacks: First, it has a single monitoring dimension, only collecting acoustic amplitude data and failing to distinguish noise sources. It is difficult to effectively differentiate between construction machinery noise and passive background noise such as traffic, daily life, and wind noise, leading to inflated monitoring data and frequent false alarms. Second, it has poor anti-interference capabilities. Equipment vibration, spatial echoes, and changes in weather conditions under complex construction site conditions can seriously affect the accuracy of acoustic acquisition, resulting in prominent data drift and distortion problems.

[0004] Some existing technologies attempt to use multi-sensor assisted monitoring, but they only achieve simple data superposition. They do not design construction status constraint mechanisms, time period dynamic segmentation strategies, and hierarchical noise decomposition logic for multi-interference scenarios on construction sites. Multi-source data suffers from problems such as temporal misalignment, feature redundancy, and scenario mismatch. They cannot distinguish between effective construction noise and passive background noise, and the monitoring accuracy and intelligence level cannot meet the current environmental monitoring standards for construction sites. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent monitoring method and system for construction site noise, aiming to solve the problems of temporal misalignment, feature redundancy, and scene mismatch in existing technologies, which make it impossible to distinguish between effective construction noise and passive background noise, and the monitoring accuracy and intelligence level cannot meet the current construction site environmental monitoring standards.

[0006] To achieve the above objectives, the present invention provides an intelligent monitoring method for construction site noise, comprising the following steps: Several monitoring nodes are set up at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors and image acquisition equipment. Multimodal raw data are collected synchronously based on a unified time reference, and all data are aligned to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals and image data. Time-frequency domain multidimensional acoustic features are extracted from noise signals, and construction state constraint information is extracted from vibration signals and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise components and background noise components. The monitoring period is dynamically divided into construction period and non-construction period. The noise signal during the non-construction period is used as the background noise reference value. Based on the difference between the total sound level during the construction period and the background noise level, the corrected equivalent sound level value of the construction noise is calculated. The noise contribution weight of each piece of construction machinery is calculated based on the noise component of construction machinery, background noise component, equivalent sound level of construction noise, meteorological signals, and image data, and the comprehensive noise emission index is output.

[0007] The process involves setting up several monitoring nodes at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. Multimodal raw data is collected synchronously based on a unified time reference, and all data is aligned to a unified time series to form a multi-source dataset. This multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. Multiple monitoring nodes are set up at preset intervals along the boundary of the construction site. The parameters of various sensor devices mounted on each node are initialized, the working conditions are adjusted and the accuracy is calibrated. A unified time reference system for the entire area is established, and the clock of the sensor devices at all monitoring nodes is synchronized. The sensor devices include noise sensors, vibration sensors, weather sensors and image acquisition devices. Multimodal raw data from the entire construction site are collected synchronously. The multimodal raw data is then subjected to time-series filtering, matching, and alignment processing to remove time-series disordered, invalid, and redundant data. All valid data are uniformly mapped to the same standard time series to construct a multi-source dataset.

[0008] Among them, the steps of extracting time-frequency domain multidimensional acoustic features from noise signals and extracting construction state constraint information from vibration signals and image data, using construction state constraint information as a priori conditions, and decomposing noise signals into construction machinery noise components and background noise components are as follows: The noise signal is subjected to global feature analysis to extract multi-dimensional acoustic core features such as time-domain fluctuation features and frequency-domain distribution features, while retaining key information such as the amplitude, frequency and fluctuation pattern of the noise signal. Vibration signals are used to extract features and identify the equipment's operating status, which includes start-up, shutdown, no-load, and operation. Analyze image data to identify scene information such as construction site workers' operations, machinery operation, and site conditions; Based on the integrated equipment operating status and scenario information, construction status constraint information is extracted.

[0009] After the step of extracting construction status constraint information from the integrated equipment operating status and scenario information: Using construction status constraint information as a priori judgment basis, we distinguish between the effective construction period and the period without construction interference at the construction site, and filter the scene status corresponding to the noise signal. By combining acoustic characteristics with the matching results of construction status, the mixed noise signal is split and separated to extract the mechanical noise component generated by the construction equipment operation and the environmental background noise component that is not related to construction.

[0010] In the step of dynamically dividing the monitoring period into construction and non-construction periods, using the noise signal from the non-construction period as the background noise reference value, and calculating the corrected equivalent sound level value of the construction noise based on the difference between the total sound level during the construction period and the background noise level: Based on the construction status identified by image data and vibration signals, the continuous monitoring period throughout the day is dynamically segmented to divide the construction period with construction work and the non-construction period without construction work. Collect all noise monitoring data during non-construction periods, statistically analyze the overall noise level and fluctuation range during these periods, and use it as a reference benchmark for site background noise. Extract the real-time total noise level data for each construction period, compare it with the background noise reference level for the corresponding period, and eliminate numerical interference and deviation caused by background noise.

[0011] In the steps of extracting the real-time total noise level data for each construction period, comparing it with the background noise reference level for the corresponding period, and eliminating numerical interference and deviation caused by background noise: The real-time total noise level data for each independent construction period is extracted one by one. This total noise level is the superimposed and mixed noise level of construction machinery noise and site background noise. The total noise level data for each construction period is compared and matched with the corresponding environment and site background noise reference level for the corresponding period to identify and remove the background noise components of the natural environment, surrounding traffic and living environment superimposed on the construction noise.

[0012] In the steps of extracting the real-time total noise level data for each construction period, comparing it with the background noise reference level for the corresponding period, and eliminating numerical interference and deviation caused by background noise: The correction process is performed based on the difference between the total sound level and the background sound level to eliminate the superposition effect of environmental background noise and obtain the corrected equivalent sound level value of construction noise.

[0013] In the step of calculating the noise contribution weight of each piece of construction machinery based on the noise component of construction machinery, the background noise component, the equivalent sound level of construction noise, meteorological signals, and image data, and outputting the comprehensive noise emission index: The noise components of the disassembled construction machinery, the background noise components, and the corrected equivalent sound level of the construction noise are collected to construct a core noise evaluation data system. By combining real-time meteorological signals, the interference of environmental factors such as temperature, humidity, wind speed, and air pressure on noise propagation and data collection is corrected. At the same time, the operating time and intensity information of each construction machine are matched with image data. By integrating various noise data, environmental correction data, and equipment operating status data, we quantitatively analyze the contribution ratio of different construction machines to the overall site noise and calculate the noise contribution weight of each construction machine.

[0014] Among these steps, after comprehensively analyzing various noise data, environmental correction data, and equipment operating status data, quantitatively analyzing the contribution ratio of different construction machinery to the overall construction site noise, and calculating the noise contribution weight corresponding to each construction machinery: By integrating multi-dimensional information such as the noise contribution weight of each piece of equipment, equivalent sound level, and environmental interference coefficient, a comprehensive noise emission index is generated to characterize the overall noise emission level of the construction site.

[0015] This invention also provides an intelligent construction site noise monitoring system, comprising a multi-source sensor data alignment module, a noise signal feature source separation module, a background noise correction module, and a noise contribution assessment module; wherein: The multi-source sensor data alignment module is used to deploy several monitoring nodes at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. It synchronously collects multimodal raw data based on a unified time reference and aligns all data to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. The noise signal feature source separation module is used to extract time-frequency domain multidimensional acoustic features from the noise signal and extract construction state constraint information from the vibration signal and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise component and background noise component. The background noise correction module is used to dynamically divide the monitoring period into construction period and non-construction period, use the noise signal of the non-construction period as the background noise reference value, and calculate the corrected equivalent sound level value of construction noise based on the difference between the total sound level and the background noise sound level during the construction period. The noise contribution assessment module is used to calculate the noise contribution weight of each construction machine based on the noise component of the construction machinery, the background noise component, the equivalent sound level value of the construction noise, meteorological signals, and image data, and output the comprehensive noise emission index.

[0016] This invention discloses an intelligent construction site noise monitoring method and system, employing a multi-source sensor data alignment module, a noise signal feature source separation module, a background noise correction module, and a noise contribution assessment module to perform the following steps: Several monitoring nodes are deployed at the construction site boundary. Each node integrates a noise sensor, a vibration sensor, a meteorological sensor, and an image acquisition device. Multimodal raw data is synchronously collected based on a unified time reference, and all data are aligned to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. Time-frequency domain multidimensional acoustic features are extracted from the noise signals, and construction state constraint information is extracted from the vibration signals and image data. Using construction status constraints as prior conditions, the noise signal is decomposed into construction machinery noise components and background noise components. The monitoring period is dynamically divided into construction periods and non-construction periods. The noise signal during non-construction periods is used as the background noise reference value. Based on the difference between the total sound level during construction periods and the background noise level, the corrected equivalent sound level value of construction noise is calculated. The noise contribution weight of each piece of construction machinery is calculated based on the construction machinery noise components, background noise components, equivalent sound level value of construction noise, meteorological signals, and image data, and a comprehensive noise emission index is output. Through the above methods, effective construction noise and passive background noise can be distinguished, improving monitoring accuracy and intelligence, thereby meeting the current environmental monitoring standards for construction sites. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of the steps of the intelligent monitoring method for construction site noise of the present invention.

[0019] Figure 2 This is a flowchart of steps S100 of the present invention.

[0020] Figure 3 This is a flowchart of steps S200 of the present invention.

[0021] Figure 4 This is a flowchart of steps S300 of the present invention.

[0022] Figure 5 This is a flowchart of steps S400 of the present invention.

[0023] Figure 6 This is a structural schematic diagram of the intelligent construction site noise monitoring system of the present invention.

[0024] Figure 7 This is a schematic diagram of the electronic device of the present invention. Detailed Implementation

[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0026] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0027] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0028] Please see Figures 1-5 This invention provides an intelligent method for monitoring construction site noise, comprising the following steps: S100: Several monitoring nodes are deployed at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. Multimodal raw data is synchronously collected based on a unified time reference, and all data are aligned to a unified time series to form a multi-source dataset. This multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. The specific process is as follows: S101: Multiple monitoring nodes are set up at preset intervals along the boundary of the construction site. The parameters of various sensor devices mounted on each node are initialized, the working conditions are adjusted and the accuracy is calibrated. A unified time reference system for the entire area is established, and the clock of the sensor devices at all monitoring nodes is synchronized. The sensor devices include noise sensors, vibration sensors, weather sensors and image acquisition devices. S102: Synchronously collect multimodal raw data from the entire construction site, perform time-series filtering, matching and alignment processing on the multimodal raw data, remove time-series disordered and invalid redundant data, and map all valid data to the same standard time series to construct a multi-source dataset.

[0029] In the above method, multiple monitoring nodes are deployed at preset intervals along the construction site boundary. Each node is installed on a bracket or fixed structure, ensuring that the noise sensor's height from the ground meets the requirements of the acoustic environment monitoring standard. Vibration sensors are tightly attached to the ground or the surface of the construction machinery base. Meteorological sensors are placed in an open, unobstructed location. The image acquisition equipment's field of view covers the main construction area. After installation, the parameters of the noise sensor, vibration sensor, meteorological sensor, and image acquisition equipment mounted on each node are initialized and configured, setting the sampling frequency, measurement range, and signal output format of each sensor. Subsequently, operational condition debugging is performed sequentially, verifying whether each sensor is in normal working condition through test signals and checking whether the signal transmission link is unobstructed. Finally, a standard sound source calibrator is used to calibrate the sound pressure level of the noise sensor, a standard vibration table is used to calibrate the sensitivity of the vibration sensor, and a standard temperature and humidity generator is used to compare and calibrate the meteorological sensor. Meanwhile, using GPS satellite clocks or IEEE 1588 precision time protocol servers as time sources, a unified time reference system covering the entire construction site is built. The clocks of the sensor devices at all monitoring nodes are synchronized and calibrated to ensure that each node and each sensor starts sampling at the same time and is marked with a unified timestamp, with the time synchronization accuracy controlled to no more than 1 millisecond.

[0030] Each monitoring node initiates synchronous data acquisition. Noise sensors continuously acquire sound pressure signals at a sampling rate of no less than 44.1 kHz, vibration sensors acquire triaxial vibration acceleration signals at a sampling rate of no less than 1 kHz, meteorological sensors acquire wind speed, wind direction, temperature, humidity, and air pressure data once per second, and image acquisition equipment acquires visible light images or video streams of the construction site at a frame rate of no less than 25 frames per second. All data are uploaded to the data center with a unified timestamp and node location identifier. After receiving the raw data from each channel, the data center first performs time-series filtering, arranging the data from each channel according to the order of timestamps and removing data with time-series errors caused by network latency or sensor frame drops. Then, it performs time-series matching on the data from different channels, associating noise, vibration, meteorological, and image data at the same moment with the timestamp as the primary key to the same record. Next, it performs alignment processing, resampling data from different sampling frequencies to a unified target time interval using interpolation methods, so that the data of all modalities are completely aligned in the time dimension. Finally, it removes invalid and redundant data, deleting outliers that exceed the reasonable range caused by abnormal fluctuations in the sensor, redundant records that have remained unchanged for a long time or have been repeatedly uploaded, and mapping all valid data to the same standard time series to construct a multi-source dataset containing time labels, spatial location labels, and modal type labels.

[0031] S200: Extracts time-frequency domain multidimensional acoustic features from noise signals and extracts construction state constraint information from vibration signals and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise components and background noise components. The specific process is as follows: S201: Perform global feature analysis on noise signals, extract multi-dimensional acoustic core features such as time-domain fluctuation features and frequency-domain distribution features, and retain key information on the amplitude, frequency, and fluctuation patterns of noise signals; S202: Perform feature mining on vibration signals to identify the equipment's operating status; the operating status includes start-up, shutdown, no-load, and operation. S203: Analyze image data to identify scene information such as construction site personnel operations, machinery operation, and site conditions; S204: Extract construction status constraint information from integrated equipment operating status and scenario information; S205: Use construction status constraint information as a priori judgment basis to distinguish between the effective construction period and the period without construction interference at the construction site, and filter the scene status corresponding to the noise signal. S206: Combining acoustic characteristics with construction status matching results, the mixed noise signal is split and separated to extract the mechanical noise component generated by construction equipment operation and the environmental background noise component that is not related to construction.

[0032] In the above method, noise signal channel data is extracted from a multi-source dataset. The continuous time-domain sound pressure signal is then segmented into frames with a preset frame length. A fixed overlap ratio is set between adjacent frames to ensure inter-frame continuity. After windowing each frame, a Fast Fourier Transform (FFT) is performed to convert the time-domain signal to the frequency domain. In the time domain, the root mean square (RMS) value of each frame is calculated as the effective sound pressure amplitude. The ratio of the peak value to the RMS value reflects the dynamic range of the signal. The rate at which the signal waveform crosses zero reflects the severity of time-domain fluctuations. Short-time energy reflects the intensity distribution of the signal within that frame. In the frequency domain, based on the frequency resolution characteristics of human hearing, the proportion of energy in each frequency band to the total energy of the entire frequency band is extracted as the spectral distribution feature. The rate of change of energy in each frequency band over time reflects the spectral fluctuation pattern. The dispersion of energy distribution in each frequency band reflects the spectral morphology of the noise. All time-domain and frequency-domain features are concatenated into a multi-dimensional acoustic core feature vector. This feature vector completely preserves the amplitude information, frequency composition, and fluctuation pattern of the noise signal.

[0033] Vibration signal channels were extracted from multi-source datasets. The raw vibration acceleration signals were bandpass filtered to remove DC offset and high-frequency electromagnetic interference components outside the target frequency band. The filtered signals were segmented by time windows. Within each time window, peak acceleration was extracted as a parameter representing mechanical impact intensity, root mean square velocity was extracted as a parameter representing vibration energy, and the dominant frequency component and its amplitude were extracted as fingerprint features of the machine's operating frequency. These vibration feature parameters were input into a pre-trained support vector machine (SVM) classifier for operational status discrimination. The training process for this classifier involved: collecting vibration sample data of various construction machinery under different operating conditions at the construction site; labeling each sample with three operational statuses: start / stop, no-load, and operation; extracting the vibration features of the sample as input; and using the labeled status labels as supervision signals to train the SVM classifier to learn the mapping relationship between vibration features and operational status. During actual monitoring, the real-time extracted vibration features were input into the trained classifier, which output whether the current operational status of the construction machinery corresponding to the vibration signal was start / stop, no-load, or operation.

[0034] Image data is extracted from multi-source datasets. After preprocessing the images, such as size normalization and brightness / contrast equalization, the data is input into a pre-trained object detection model for scene information recognition. The training process of this object detection model is as follows: a large number of labeled images of construction sites are collected. The labels include the bounding boxes and categories of construction machinery, the detection boxes of workers, material stacking areas, and construction activity areas. Using the labeled images as training samples, supervised training is performed using YOLO or Faster R-CNN as the backbone network. Backpropagation is used to optimize the network parameters so that the model has the ability to detect and classify construction site scene elements from images. During actual monitoring, the model infers for each frame of image and outputs the category labels and location information of various types of construction machinery, the distribution of workers, and the active areas and ranges of construction activities within the site. By tracking the detection results of consecutive frames in time, the model further identifies the pose changes and movement trajectories of each construction machine to determine whether it is in operation.

[0035] The operating status labels of each construction machine output in step S202 are spatiotemporally correlated and fused with the construction machine category labels, worker distribution information and construction activity area information output in step S203. Using the monitoring node location as a spatial reference, the vibration operation status near each node is cross-validated with the corresponding image recognition results. When the outputs of the two modes point to the same construction status, a high confidence level is assigned. When there is a conflict, the image recognition result is the main factor and the vibration recognition result is the auxiliary factor for comprehensive judgment. Finally, construction status constraint information is formed that covers whether there is construction activity in all areas of the construction site, which types of construction machines are operating, what stage of operation each machine is currently in, and the spatial distribution range of construction activities.

[0036] Using the construction state constraint information extracted in step S204 as a priori criterion, the noise signals in the multi-source dataset are labeled with a time axis. Specifically, the construction state constraint information at each moment is associated with the noise signal record at that moment, and the scene state type corresponding to each moment is labeled, including the state with large machinery operation, the state with small machinery operation, the state with only manual operation, and the state with no construction interference. Based on the above scene state labeling, all monitoring periods are divided into effective construction periods and periods without construction interference. The effective construction period is the period with machinery or manual operation, and the period without construction interference is the period when the entire construction site is quiet and only environmental background sound exists. At the same time, a corresponding scene state label is matched for each noise signal to clarify the construction conditions under which each noise signal was collected.

[0037] Based on the noise signal and scene state matching results obtained in step S205, the mixed noise signal is split and separated. First, all noise signals belonging to the non-construction interference period are aggregated into the environmental background noise base. Then, during the effective construction period, based on the operating status of each construction machine identified in step S202 and the machine category identified in step S203, an acoustic feature template is established for each type of construction machine. This template includes the typical octave spectrum shape and time-domain envelope characteristics of this type of machine during the operation period. Next, the constrained independent component analysis algorithm is used to separate the mixed noise signal. This algorithm uses the acoustic feature template of each machine as the reference signal constraint and the active period of the machine determined by the vibration signal as the amplitude activity constraint. In the iterative solution process, the separated source signals satisfy the statistical independence criterion and match the corresponding reference feature template and activity constraint. Finally, the mixed noise signal is peeled off layer by layer, decomposing the machine noise components independently generated by each type of construction machine and the environmental background noise components unrelated to construction.

[0038] S300: The monitoring period is dynamically divided into construction and non-construction periods. The noise signal during the non-construction period is used as the background noise reference value. Based on the difference between the total sound level during the construction period and the background noise level, the corrected equivalent sound level value of the construction noise is calculated. The specific process is as follows: S301: Based on the construction status identified by image data and vibration signals, the continuous monitoring period throughout the day is dynamically segmented to divide the construction period with construction work and the non-construction period without construction work. S302: Collect all noise monitoring data during non-construction periods, statistically analyze the overall noise level and fluctuation range during these periods, and use it as a reference benchmark for site background noise. S303: Extract the real-time total noise level data for each construction period, compare it with the background noise reference level for the corresponding period, and eliminate numerical interference and deviation caused by background noise. S304: Based on the difference between the total sound level and the background sound level, the correction process is performed to eliminate the superposition effect of environmental background noise and obtain the corrected equivalent sound level value of construction noise.

[0039] In the above method, based on the construction state constraint information extracted in step S204 and the scene state matching result in step S205, the continuous monitoring period throughout the day is dynamically segmented moment by moment. Using the existence marker of construction activity and the scene state label in the construction state constraint information as the primary criteria, supplemented by whether the peak acceleration of the vibration signal exceeds a preset start / stop threshold as an auxiliary criterion, each moment is marked as either an active construction state or a static construction state. Time intervals continuously marked as active construction states are merged into an independent construction period, and time intervals continuously marked as static construction states are merged into a non-construction period. Finally, the monitoring data throughout the day is segmented into alternating sequences of construction periods and non-construction periods, with each period recording its start time, end time, and duration.

[0040] Noise monitoring data for all non-construction periods identified in step S301 are collected. The instantaneous sound level data for each non-construction period are averaged using an energy-averaging method to calculate the equivalent sound level for that period, which serves as the representative value of the background noise for that period. Considering the potential differences in ambient background noise across different times of the day, the equivalent sound levels for all non-construction periods are statistically analyzed to calculate their average value, standard deviation, and fluctuation range. If the differences in background sound levels between different non-construction periods are within acceptable limits, the overall average value is taken as the fixed reference benchmark for the site's background noise for that day. If there are significant diurnal differences in background sound levels or if certain periods are abnormally high due to unforeseen external events, outliers are marked and removed. Time-matched background noise reference benchmarks are established for different time periods to ensure that the reference benchmarks correspond to the actual environmental conditions.

[0041] The real-time total noise level data for each independent construction period segment identified in step S301 is extracted one by one. This total noise level is the mixed sound level resulting from the superposition of construction noise generated by construction machinery and background noise. For each construction period segment, the total equivalent sound level for that period is first calculated by averaging the instantaneous sound levels at all sampling times within that period. Then, based on the time location of that construction period, the background noise reference sound level value for the corresponding time segment is matched with the background noise reference benchmark established in step S302. The total equivalent sound level is compared with the background reference sound level item by item, and the difference between the two is calculated. The relative strength of background components superimposed on the construction noise, such as natural environmental wind and rain noise, surrounding road traffic noise, and residential noise, is identified to complete the matching and verification of the total sound level of the construction period with the background reference sound level.

[0042] Based on the difference between the total equivalent sound level calculated in step S303 and the background reference sound level, a graded correction process is performed to eliminate the superposition effect of environmental background noise. When the difference between the total sound level and the background sound level is greater than a preset first threshold, it is determined that the background noise is much lower than the construction noise, and its impact on the total sound level can be ignored. The total sound level is directly used as the corrected equivalent sound level value of the construction noise. When the difference is in the middle threshold range, the corresponding correction deduction is determined according to the specific range of the difference. The smaller the difference, the greater the relative impact of the background noise, and the correction deduction is increased accordingly. The corrected equivalent sound level value of the construction noise is obtained by deducting the correction from the total sound level. When the difference is less than a preset minimum threshold, it is determined that the construction noise is submerged by the background noise. The reliability of the measurement results for this period is too low. The data for this period is marked as invalid data, no correction value is output, and it is excluded in subsequent evaluations. The corresponding equivalent sound level values ​​of construction noise are output after correction for all valid construction periods.

[0043] S400: Calculates the noise contribution weight of each piece of construction machinery based on the noise components of construction machinery, background noise components, equivalent sound level of construction noise, meteorological signals, and image data, and outputs a comprehensive noise emission index; the specific process is as follows: S401: Collect the noise components of the construction machinery after dismantling, the background noise components, and the corrected equivalent sound level of the construction noise to construct a core noise evaluation data system. S402: Combines real-time meteorological signals to correct the interference of environmental factors such as temperature, humidity, wind speed, and air pressure on noise propagation and data collection, and combines image data to match the working time and intensity information of each construction machine. S403: By integrating various noise data, environmental correction data, and equipment operating status data, we can quantitatively analyze the contribution ratio of different construction machinery to the overall construction site noise and calculate the noise contribution weight of each construction machinery. S404: Integrates multiple dimensions of information, including the noise contribution weight of each piece of equipment, equivalent sound level, and environmental interference coefficient, to generate a comprehensive noise emission index that characterizes the overall noise emission level of the construction site.

[0044] In the above method, the time series waveforms and average power intensity of the various types of construction machinery noise components obtained by decomposition in step S206, the time series and average power level of the environmental background noise components separated in step S206, and the equivalent sound level values ​​of construction noise for each construction period that have been corrected and marked as valid in step S304 are collected. The above three types of noise data are uniformly indexed and associated according to the time axis to construct a core noise evaluation dataset containing labels for machinery noise components, background noise components, and corrected sound level labels, which serves as the basic data system for subsequent multi-dimensional evaluation.

[0045] Meteorological signal data corresponding to the monitoring period are extracted from multi-source datasets, including wind speed, wind direction, temperature, humidity, and air pressure. The impact of these meteorological factors on noise acquisition and propagation is corrected one by one. Specifically, the attenuation correction amount along the sound propagation path is calculated based on wind speed and direction data. Temperature changes cause changes in air density, thus affecting sound speed and acoustic impedance. Humidity and air pressure changes also affect sound absorption and attenuation in the atmosphere. The correction amounts of the above meteorological factors are superimposed according to acoustic propagation theory to obtain a comprehensive environmental interference correction coefficient. Simultaneously, the types and quantities of construction machinery appearing in each construction period are matched from the image recognition results of step S203. The operating time of each type of machinery is calculated based on the spatiotemporal changes in machinery positions in consecutive frames. The work intensity level is determined based on the density of machinery movement trajectories and the coverage of the work area, forming information on the operating time and work intensity of each type of machinery.

[0046] Based on the noise core evaluation data constructed in step S401, the environmental interference correction coefficient obtained in step S402, and the operating time and intensity information of various types of machinery, a quantitative analysis of the noise contribution of each type of construction machinery is performed. First, the independent noise radiation intensity of each type of machinery is extracted from the machinery noise components. Then, based on the actual operating time and intensity level of that type of machinery within the evaluation period, its effective noise energy contribution is calculated. The noise contribution of machinery with short operating times or low operating intensities is reduced accordingly, while the noise contribution of machinery with long operating times or high operating intensities is amplified accordingly. The environmental interference correction coefficient is applied as a global scaling factor to the noise contribution value of all machinery to eliminate the systematic bias caused by meteorological conditions in sound level measurement. After the above correction and weighting, the noise contribution values ​​of all construction machinery are normalized so that the sum of the contribution values ​​of each type of machinery is 100%, thus obtaining the contribution weight of each construction machinery in the overall noise emission of the construction site.

[0047] The noise contribution weights of each piece of equipment calculated in step S403, the equivalent sound level values ​​of construction noise for each time period corrected in step S304, the environmental interference correction coefficient obtained in step S402, and other multi-dimensional information such as the proportion of construction periods, frequency of exceeding standards, and duration of exceeding standards are integrated using a weighted aggregation method. Specifically, the corrected equivalent sound level value of construction noise is used as the basic acoustic index, the contribution weights of each piece of equipment are used as the distribution structure information of various sound sources, the environmental interference correction coefficient is used as the measurement reliability factor, and the frequency and duration of exceeding standards are used as the time dimension influence factors. A comprehensive score ranging from zero to the upper limit is calculated through multi-dimensional weighted fusion. This score is the comprehensive noise emission index, which characterizes the overall noise emission level of the construction site. At the same time, based on the numerical range of the comprehensive index, four noise risk levels—low risk, medium risk, high risk, and severe exceeding standards—are determined. The contribution weights of each piece of construction machinery, the comprehensive noise emission index, and the risk level are output as the final evaluation results.

[0048] Corresponding to the aforementioned embodiments of the intelligent monitoring method for construction site noise, this application also provides embodiments of an intelligent monitoring system for construction site noise.

[0049] Figure 6 This is a structural schematic diagram of an intelligent construction site noise monitoring system according to an exemplary embodiment. (Refer to...) Figure 6 The system may include: a multi-source sensor data alignment module, a noise signal feature source separation module, a background noise correction module, and a noise contribution assessment module; wherein: The multi-source sensor data alignment module is used to deploy several monitoring nodes at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. It synchronously collects multimodal raw data based on a unified time reference and aligns all data to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. The noise signal feature source separation module is used to extract time-frequency domain multidimensional acoustic features from the noise signal and extract construction state constraint information from the vibration signal and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise component and background noise component. The background noise correction module is used to dynamically divide the monitoring period into construction period and non-construction period, use the noise signal of the non-construction period as the background noise reference value, and calculate the corrected equivalent sound level value of construction noise based on the difference between the total sound level and the background noise sound level during the construction period. The noise contribution assessment module is used to calculate the noise contribution weight of each construction machine based on the noise component of the construction machinery, the background noise component, the equivalent sound level value of the construction noise, meteorological signals, and image data, and output the comprehensive noise emission index.

[0050] In this embodiment, the multi-source sensor data alignment module deploys several monitoring nodes at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. It synchronously collects multimodal raw data based on a unified time reference and aligns all data to a unified time series, forming a multi-source dataset. This multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. The noise signal feature source separation module extracts time-frequency domain multidimensional acoustic features from the noise signal and extracts construction state constraint information from the vibration signal and image data. Using the construction state constraint information as a priori conditions, it decomposes the noise signal into construction machinery noise components. The monitoring system includes a background noise component; the background noise correction module dynamically divides the monitoring period into construction and non-construction periods, using the noise signal from the non-construction period as a background noise reference value, and calculates the corrected equivalent sound level value of construction noise based on the difference between the total sound level and the background noise level during the construction period; the noise contribution assessment module calculates the noise contribution weight of each piece of construction machinery based on the construction machinery noise component, background noise component, equivalent sound level value of construction noise, meteorological signals, and image data, and outputs a comprehensive noise emission index; through the above methods, effective construction noise and passive background noise can be distinguished, improving monitoring accuracy and intelligence, thereby meeting current construction site environmental monitoring standards.

[0051] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0052] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0053] Accordingly, this application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; and, when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the intelligent construction site noise monitoring method described above. Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities within a construction site noise intelligent monitoring system provided in an embodiment of the present invention, except... Figure 7 In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0054] Accordingly, this application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the intelligent construction site noise monitoring method described above. The computer-readable storage medium can be an internal storage unit of any data processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0055] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0056] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method for intelligent monitoring of construction site noise, characterized in that, Includes the following steps: Several monitoring nodes are set up at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors and image acquisition equipment. Multimodal raw data are collected synchronously based on a unified time reference, and all data are aligned to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals and image data. Time-frequency domain multidimensional acoustic features are extracted from noise signals, and construction state constraint information is extracted from vibration signals and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise components and background noise components. The monitoring period is dynamically divided into construction period and non-construction period. The noise signal during the non-construction period is used as the background noise reference value. Based on the difference between the total sound level during the construction period and the background noise level, the corrected equivalent sound level value of the construction noise is calculated. The noise contribution weight of each piece of construction machinery is calculated based on the noise component of construction machinery, background noise component, equivalent sound level of construction noise, meteorological signals, and image data, and the comprehensive noise emission index is output.

2. The intelligent monitoring method for construction site noise as described in claim 1, characterized in that, Several monitoring nodes are deployed at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. Multimodal raw data are synchronously collected based on a unified time reference, and all data are aligned to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. Multiple monitoring nodes are set up at preset intervals along the boundary of the construction site. The parameters of various sensor devices mounted on each node are initialized, the working conditions are adjusted and the accuracy is calibrated. A unified time reference system for the entire area is established, and the clock of the sensor devices at all monitoring nodes is synchronized. The sensor devices include noise sensors, vibration sensors, weather sensors and image acquisition devices. Multimodal raw data from the entire construction site are collected synchronously. The multimodal raw data is then subjected to time-series filtering, matching, and alignment processing to remove time-series disordered, invalid, and redundant data. All valid data are uniformly mapped to the same standard time series to construct a multi-source dataset.

3. The intelligent construction site noise monitoring method as described in claim 1, characterized in that, In the steps of extracting time-frequency domain multidimensional acoustic features from noise signals and extracting construction state constraint information from vibration signals and image data, using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise components and background noise components: The noise signal is subjected to global feature analysis to extract multi-dimensional acoustic core features such as time-domain fluctuation features and frequency-domain distribution features, while retaining key information such as the amplitude, frequency and fluctuation pattern of the noise signal. Vibration signals are used to extract features and identify the equipment's operating status, which includes start-up, shutdown, no-load, and operation. Analyze image data to identify scene information such as construction site workers' operations, machinery operation, and site conditions; Based on the integrated equipment operating status and scenario information, construction status constraint information is extracted.

4. The intelligent monitoring method for construction site noise as described in claim 3, characterized in that, After the step of extracting construction status constraint information based on integrated equipment operating status and scenario information: Using construction status constraint information as a priori judgment basis, we distinguish between the effective construction period and the period without construction interference at the construction site, and filter the scene status corresponding to the noise signal. By combining acoustic characteristics with the matching results of construction status, the mixed noise signal is split and separated to extract the mechanical noise component generated by the construction equipment operation and the environmental background noise component that is not related to construction.

5. The intelligent construction site noise monitoring method as described in claim 1, characterized in that, In the step of dynamically dividing the monitoring period into construction and non-construction periods, using the noise signal from the non-construction period as the background noise reference value, and calculating the corrected equivalent sound level value of construction noise based on the difference between the total sound level during the construction period and the background noise level: Based on the construction status identified by image data and vibration signals, the continuous monitoring period throughout the day is dynamically segmented to divide the construction period with construction work and the non-construction period without construction work. Collect all noise monitoring data during non-construction periods, statistically analyze the overall noise level and fluctuation range during these periods, and use it as a reference benchmark for site background noise. Extract the real-time total noise level data for each construction period, compare it with the background noise reference level for the corresponding period, and eliminate numerical interference and deviation caused by background noise.

6. The intelligent construction site noise monitoring method as described in claim 5, characterized in that, In the steps of extracting the real-time total noise level data for each construction period, comparing it with the background noise reference level for the corresponding period, and eliminating numerical interference and deviation caused by background noise: The real-time total noise level data for each independent construction period is extracted one by one. This total noise level is the superimposed and mixed noise level of construction machinery noise and site background noise. The total noise level data for each construction period is compared and matched with the corresponding environment and site background noise reference level for the corresponding period to identify and remove the background noise components of the natural environment, surrounding traffic and living environment superimposed on the construction noise.

7. The intelligent monitoring method for construction site noise as described in claim 5, characterized in that, In the steps of extracting the real-time total noise level data for each construction period, comparing it with the background noise reference level for the corresponding period, and eliminating numerical interference and deviation caused by background noise: The correction process is performed based on the difference between the total sound level and the background sound level to eliminate the superposition effect of environmental background noise and obtain the corrected equivalent sound level value of construction noise.

8. The intelligent monitoring method for construction site noise as described in claim 1, characterized in that, In the steps of calculating the noise contribution weight of each piece of construction machinery based on the noise component of construction machinery, background noise component, equivalent sound level of construction noise, meteorological signals, and image data, and outputting the comprehensive noise emission index: The noise components of the disassembled construction machinery, the background noise components, and the corrected equivalent sound level of the construction noise are collected to construct a core noise evaluation data system. By combining real-time meteorological signals, the interference of environmental factors such as temperature, humidity, wind speed, and air pressure on noise propagation and data collection is corrected. At the same time, the operating time and intensity information of each construction machine are matched with image data. By integrating various noise data, environmental correction data, and equipment operating status data, we quantitatively analyze the contribution ratio of different construction machines to the overall site noise and calculate the noise contribution weight of each construction machine.

9. The intelligent monitoring method for construction site noise as described in claim 8, characterized in that, After comprehensively analyzing various noise data, environmental correction data, and equipment operating status data, quantitatively analyzing the contribution ratio of different construction machinery to the overall construction site noise, and calculating the noise contribution weight of each construction machinery: By integrating multi-dimensional information such as the noise contribution weight of each piece of equipment, equivalent sound level, and environmental interference coefficient, a comprehensive noise emission index is generated to characterize the overall noise emission level of the construction site.

10. A construction site noise intelligent monitoring system, employing the construction site noise intelligent monitoring method as described in claim 1, characterized in that, It includes a multi-source sensor data alignment module, a noise signal feature source separation module, a background noise correction module, and a noise contribution assessment module; among which: The multi-source sensor data alignment module is used to deploy several monitoring nodes at the construction site boundary. Each node integrates noise sensors, vibration sensors, meteorological sensors, and image acquisition equipment. It synchronously collects multimodal raw data based on a unified time reference and aligns all data to a unified time series to form a multi-source dataset. The multi-source dataset includes noise signals, meteorological signals, vibration signals, and image data. The noise signal feature source separation module is used to extract time-frequency domain multidimensional acoustic features from the noise signal and extract construction state constraint information from the vibration signal and image data. Using the construction state constraint information as a priori conditions, the noise signal is decomposed into construction machinery noise component and background noise component. The background noise correction module is used to dynamically divide the monitoring period into construction period and non-construction period, use the noise signal of the non-construction period as the background noise reference value, and calculate the corrected equivalent sound level value of construction noise based on the difference between the total sound level and the background noise sound level during the construction period. The noise contribution assessment module is used to calculate the noise contribution weight of each construction machine based on the noise component of the construction machinery, the background noise component, the equivalent sound level value of the construction noise, meteorological signals, and image data, and output the comprehensive noise emission index.

Images