An intelligent safety inspection system for construction sites

Abstract

The present invention discloses an intelligent safety inspection system for construction sites, which relates to the technical field of construction inspection. A data acquisition module is used to obtain multiple data sources of the construction site. The data fusion and processing module dynamically optimizes heterogeneous data through time synchronization, format standardization and noise filtering in combination with an adaptive weighted fusion algorithm. The intelligent analysis and early warning module evaluates the degree of data difference based on machine learning, uses cluster analysis to mine abnormal patterns for high-difference data, identifies safety risks for low-difference data, and dynamically adjusts early warning strategies. The terminal interaction module pushes inspection results and early warning information to a mobile APP, a monitoring center display screen, and a voice broadcast device, so that management personnel can remotely view the construction status and make quick decisions. The present invention can significantly improve the intelligence level of safety inspections on construction sites, realize efficient and accurate safety hazard identification and dynamic early warning, and effectively reduce the risk of safety accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of construction inspection, and in particular to an intelligent safety inspection system for a construction site. Background Art

[0002] With the rapid development of the construction industry, construction site safety management has become increasingly important. Traditional safety inspections rely primarily on manual inspections and regular patrols, but this approach presents numerous challenges, such as low inspection frequency, subjectivity, and delayed detection of potential hazards. In recent years, the rapid development of intelligent technologies has provided new insights for construction site safety inspections. By integrating technologies such as the Internet of Things, artificial intelligence, and big data analytics, real-time monitoring and intelligent management of construction sites can be achieved, thereby improving the efficiency and accuracy of safety inspections.

[0003] Existing technologies have the following shortcomings: In intelligent safety inspection systems, construction sites can present a variety of safety risks, including illegal human operation, falling objects, and mechanical equipment failures. These risks require monitoring through multiple data sources, including video surveillance, sensors, and RFID. However, these data sources vary significantly in information format, sampling frequency, and transmission delay. Efficiently integrating and analyzing this heterogeneous data in a noisy and dynamically changing construction environment to ensure the real-time and accuracy of the inspection system remains a pressing technical challenge. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent safety inspection system for construction sites to solve the shortcomings of the background technology.

[0005] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: an intelligent safety inspection system for a construction site, comprising a data acquisition module, a data fusion and processing module, an intelligent analysis and early warning module, and a terminal interaction module;

[0006] A data acquisition module is used to acquire multi-source data from the construction site, including video surveillance data, sensor data, RFID identification information, and environmental data. The sensor data includes temperature, humidity, dust concentration, and vibration intensity data.

[0007] The data fusion and processing module is used to synchronize the time of the collected multi-source data, standardize the data format, and filter noise, and dynamically optimize and fuse different data sources based on an adaptive weighted fusion algorithm;

[0008] The intelligent analysis and early warning module is used to evaluate the degree of difference of fused multi-source data based on machine learning models. For multi-source data with high degree of difference, cluster analysis is used to further explore abnormal patterns. For multi-source data with low degree of difference, security risks are identified and early warning strategies of different levels are dynamically adjusted based on the risk level and the degree of difference of multi-source data, thus achieving real-time security warning.

[0009] The terminal interaction module is used to push safety inspection results and warning information to the management terminal, including mobile APP, monitoring center display screen and voice broadcast equipment. Management personnel can remotely view the construction site conditions and make safety decisions through the terminal interaction module.

[0010] Preferably, in the intelligent analysis and early warning module, the vibration frequency change rate is analyzed by fast Fourier transform to generate a vibration frequency change rate fluctuation index, and the vibration frequency change rate fluctuation index is obtained by:

[0011] A high-precision acceleration sensor is used to collect the vibration signal of the construction structure in time series. The collected time domain vibration signal A(t) is transformed by FFT and converted into the frequency domain to obtain the spectrum A(f): ; The amplitude spectrum calculated by FFT: ;in, is the kth frequency component, k=0,1,2,...,N / 2, They are Find the real and imaginary parts of the frequency component with the largest amplitude in the FFT result, that is, the main vibration frequency : ; Compare the changes in the main vibration frequency in adjacent time windows and calculate the vibration frequency change rate , the expression is: ;in: is the main vibration frequency of the i-th time window, is the main vibration frequency of the i-1th time window, Δt is the time window interval; calculate the vibration frequency change rate fluctuation index VFVFI, the expression is: ; Where: M is the total number of time windows in the calculation period, is the mean of the vibration frequency change rate in all windows.

[0012] Preferably, the LSTM model is used to analyze the changes in the stability of the construction workers' movement patterns to generate a movement stability anomaly index. The method for obtaining the movement stability anomaly index is as follows:

[0013] The inertial sensor is used to sample the motion state of the construction workers. The sensor data at each moment is expressed as: ;in: is the triaxial acceleration, is the three-axis angular velocity, is the three-axis magnetometer data;

[0014] LSTM network training includes: input layer: input data is shape; LSTM layer: extract time series features and output hidden state ; Fully connected layer: mapped to stability score ; Loss function: Use mean square error to calculate the error between the LSTM prediction value and the true stability score to optimize the model;

[0015] Use the LSTM model to predict the stability score within each time window H Then, calculate the change rate of motion mode stability , the expression is: ;in: is the stability score of the current window, is the stability score of the previous window, Δt is the window interval; the motion stability abnormality index MSAI is calculated using the following formula: ; Where: Q is the total number of time windows, is the stability change rate of the i-th window, is the mean of the stability change rate in all windows.

[0016] Preferably, the vibration frequency change rate fluctuation index and the motion stability anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as inputs of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the difference degree value labels of the fused multi-source data as the prediction target, and uses minimizing the sum of the prediction errors of the difference degree value labels of all fused multi-source data as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The difference degree value of the fused multi-source data is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0017] Preferably, the obtained difference degree value of the fused multi-source data is compared with a difference degree reference threshold value pre-set based on historical data. If the difference degree value of the fused multi-source data is greater than or equal to the pre-set difference degree reference threshold value, it indicates that the difference degree of the fused multi-source data is high, and it is marked as multi-source data with high difference degree; if the difference degree value of the fused multi-source data is less than the pre-set difference degree reference threshold value, it indicates that the difference degree of the fused multi-source data is low, and it is marked as multi-source data with low difference degree.

[0018] Preferably, for multi-source data with high degree of difference, cluster analysis is used to further mine abnormal patterns. Specifically, the input data for cluster analysis comes from multi-source data samples with high degree of difference, and each sample contains multiple eigenvalues: ; Wherein: VFVFI is the vibration frequency change rate fluctuation index, MSAI is the motion stability abnormality index; The main vibration frequency, is the rate of change of vibration frequency, is the standard deviation of the vibration change rate, Score the stability of LSTM predictions, is the rate of change of motion stability; all samples constitute a high-dimensional feature dataset: , J is the total number of features; select the K value, initialize K cluster centers, and iteratively optimize: ;in, is the center of the k-th category, is the set of samples belonging to this category; calculate the Euclidean distance from each data point to the cluster center , and classify it into the nearest cluster: ; represents the jth eigenvalue of the i-th sample; Represents the value of the kth cluster center on the jth feature dimension; it is updated iteratively until convergence; high-variance data at the construction site are clustered into categories such as abnormal equipment vibration, unstable construction structure, and abnormal personnel behavior.

[0019] Preferably, based on the risk level and the degree of difference of multi-source data, different levels of early warning strategies are dynamically adjusted to achieve real-time security early warning, specifically:

[0020] The risk level and the difference between multi-source data are used as the input of fuzzy logic, and the early warning strategies of different levels are used as the output of fuzzy logic, and dynamic regulation is carried out through fuzzy logic.

[0021] The fuzzy membership function is used to fuzzify the input, the fuzzy rule base is used for reasoning, and the center of gravity method is used to defuzzify and calculate the final warning level value. , ;in: is the membership value corresponding to different rules, Numerical mapping for different early warning strategies;

[0022] According to the results of fuzzy reasoning, corresponding early warning measures are dynamically executed.

[0023] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0024] 1. This invention uses a data acquisition module to systematically integrate multiple data sources, including video surveillance, sensors, and RFID, ensuring comprehensive awareness of the construction site. The data fusion and processing module utilizes an adaptive weighted fusion algorithm to address differences in format, sampling frequency, and transmission delay among different data sources, improving data consistency and reliability. The intelligent analysis and early warning module, based on machine learning and polynomial regression models, accurately assesses the degree of data discrepancy by calculating the Vibration Frequency Change Fluctuation Index (VFVFI) and the Motion Stability Anomaly Index (MSAI). Cluster analysis is then used to further identify patterns such as abnormal equipment vibration, unstable construction structures, and abnormal personnel behavior, enabling efficient and accurate risk identification and graded early warning.

[0025] 2. The present invention realizes dynamic early warning adjustment under different risk levels through fuzzy logic reasoning, avoiding the limitations of traditional fixed threshold early warning methods, enabling the system to intelligently adjust the early warning level according to the actual situation on site, and improving the accuracy and response speed of early warning. The terminal interaction module supports mobile APP, monitoring center display screen and voice broadcast. Managers can remotely view inspection results and early warning information in real time, make safety decisions, and improve management efficiency. The present invention can significantly improve the intelligence, precision and real-time nature of construction safety monitoring, reduce the risk of construction accidents, improve the overall safety management level of the construction site, and assist in the construction of intelligent construction sites. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0027] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0029] For examples, see Figure 1 As shown, the intelligent safety inspection system for a construction site described in this embodiment includes a data acquisition module, a data fusion and processing module, an intelligent analysis and early warning module, and a terminal interaction module;

[0030] A data acquisition module is used to acquire multi-source data from the construction site, including video surveillance data, sensor data, RFID identification information, and environmental data. The sensor data includes temperature, humidity, dust concentration, and vibration intensity data.

[0031] The data fusion and processing module is used to synchronize the time of the collected multi-source data, standardize the data format, and filter noise, and dynamically optimize and fuse different data sources based on an adaptive weighted fusion algorithm;

[0032] The intelligent analysis and early warning module is used to evaluate the degree of difference of fused multi-source data based on machine learning models. For multi-source data with high differences, cluster analysis is used to further explore abnormal patterns. For multi-source data with low differences, security risks are identified and early warning strategies of different levels are dynamically adjusted based on the risk level and the changes in the degree of difference of multi-source data, achieving real-time security warnings.

[0033] The terminal interaction module is used to push safety inspection results and warning information to the management terminal, including mobile APP, monitoring center display screen and voice broadcast equipment. Management personnel can remotely view the construction site conditions and make safety decisions through the terminal interaction module.

[0034] The data acquisition module is responsible for acquiring various data from the construction site to comprehensively understand the construction environment, personnel, and equipment status. Video surveillance data is collected in real time by high-definition cameras deployed in key areas. These cameras not only provide high-definition video streams but also support night vision and zoom capabilities to accurately capture construction site details in all lighting conditions. Video data can be used to detect hazardous scenarios such as illegal human operation, abnormal equipment operation, and high-altitude work.

[0035] Sensor data includes environmental variables such as temperature, humidity, dust concentration, and vibration intensity. Temperature and humidity sensors are primarily installed at various heights and locations within the construction area to monitor environmental changes in real time and prevent extreme weather from impacting construction safety. Dust concentration sensors are used to detect the amount of dust particles in the air, particularly in high-dust work areas such as welding, cutting, and grinding. This allows for timely identification of air quality violations and reduces the risk of workers inhaling harmful substances. Vibration intensity sensors are typically installed on building structures, scaffolding, and mechanical equipment to monitor vibration during construction and prevent structural instability or equipment failure due to excessive vibration.

[0036] RFID identification information is used to track personnel and equipment on construction sites. Construction workers wear RFID tags. When they enter or leave designated areas, RFID readers automatically record their entry and exit information. This information, combined with the attendance management system, ensures that construction workers are working in the correct areas within the specified timeframe. RFID technology can also be used for the intelligent identification and positioning of key equipment, such as large machinery like cranes and excavators. RFID tags record the equipment's operating status, maintenance time, and user information, enhancing the intelligence of equipment management.

[0037] Environmental data primarily includes information such as wind speed and noise. Wind speed sensors are often installed in high-altitude work areas, such as on tower cranes and scaffolding, to monitor wind intensity in real time. When wind speeds exceed safety standards, the system automatically issues an alert, reminding construction workers to suspend high-altitude work. Noise sensors are used to monitor noise levels at construction sites, particularly near residential areas, to control noise pollution and minimize environmental impact.

[0038] Through the multi-source data acquisition capability of the data acquisition module, the system can perceive the overall status of the construction site in real time, provide high-quality information input for subsequent data fusion and intelligent analysis, and ensure the comprehensiveness and accuracy of safety inspections.

[0039] The data fusion and processing module is used to deeply process multi-source data collected at the construction site to ensure data accuracy, timeliness, and consistency, providing high-quality data support for subsequent intelligent analysis. This module primarily includes time synchronization, data format standardization, noise filtering, and data optimization and fusion based on an adaptive weighted fusion algorithm.

[0040] First, time synchronization is a key step in ensuring the effective fusion of multi-source data. The sampling frequencies and timestamps of various sensors, video surveillance equipment, and RFID readers at the construction site may differ. Directly using these sensors can lead to data asynchrony, affecting the accuracy of analysis results. To address this issue, the system employs a global time synchronization mechanism, using Network Time Protocol (NTP) or GPS timing technology to ensure that the timestamps of all data sources are consistent with the system time. Furthermore, for sensors with clock drift, the system regularly performs time calibration and uses interpolation algorithms to time-align the data during the data processing phase to ensure that the data is analyzed based on the same time reference.

[0041] Secondly, data format standardization addresses the issue of inconsistent formats across different data sources. Construction sites host a wide variety of data types. For example, video data is typically in streaming format, sensor data may be discrete values, and RFID information consists of an identifier and a time record. Through a unified data interface, the system converts all data into a standardized format, such as JSON or a database storage format. This allows different types of data to be processed and stored on the same platform. Furthermore, for video data, the system performs keyframe extraction and feature encoding, converting the video information into structured data for correlation analysis with sensor data.

[0042] Noise filtering is a crucial step in data processing, especially in complex environments like construction sites, where data can be affected by electromagnetic interference, environmental changes, or equipment errors. The system employs a variety of data cleaning and denoising algorithms. For example, the Kalman filter algorithm is used on sensor data to smooth out sudden changes and improve data stability. For video image data, a convolutional neural network (CNN) is used for denoising, filtering out interference caused by light variations or dust obstructions. For RFID signals, average filtering and outlier detection are used to eliminate misreadings caused by signal reflections or obstructions.

[0043] Finally, the system dynamically optimizes and fuses different data sources using an adaptive weighted fusion algorithm to improve data reliability and integrity. This algorithm dynamically adjusts the weights of different data sources based on their credibility, historical error rates, and environmental adaptability. For example, in well-lit conditions, video surveillance data is given a higher weight, while at night or in inclement weather, RFID and sensor data are given a higher weight. Furthermore, the system integrates historical data and continuously optimizes weight distribution through machine learning models, making the data fusion results more accurate and stable.

[0044] Through the data fusion and processing module, the system can effectively solve problems such as asynchrony, format mismatch and noise interference caused by multi-source heterogeneous data at the construction site, providing high-quality input data for intelligent analysis and early warning, thereby improving the accuracy and real-time performance of safety inspections.

[0045] The intelligent analysis and early warning module is used to evaluate the degree of difference in the fused multi-source data based on machine learning models, including:

[0046] The vibration frequency change rate is a key parameter for structural safety at construction sites, primarily used to monitor the stability of buildings, scaffolding, or large equipment. Unlike traditional vibration intensity testing, the vibration frequency change rate focuses on the dynamic changes in vibration patterns, such as mode drift caused by uneven loading, mechanical shock, or structural looseness. This parameter can be acquired using high-precision accelerometers or microelectromechanical systems (MEMS) sensors and analyzed using signal processing techniques such as fast Fourier transforms (FFTs).

[0047] When the vibration frequency change rate remains stable or fluctuates within a preset safety range, it usually indicates that the construction structure or equipment is operating normally.

[0048] When the rate of change of vibration frequency increases abnormally (for example, the change exceeds a set threshold in a short period of time), it may mean that the scaffolding is loose, the foundation is settling, or even an early sign of equipment failure.

[0049] When the rate of change of vibration frequency decreases but the amplitude increases, it may indicate nonlinear deformation of the structure, such as crack growth in concrete or fatigue damage in steel structures.

[0050] The vibration frequency change rate is analyzed by fast Fourier transform to generate a vibration frequency change rate fluctuation index, wherein the vibration frequency change rate fluctuation index is obtained as follows:

[0051] Use a high-precision acceleration sensor to collect the vibration signal a(t) of the construction structure in time series and preprocess the data (denoising and normalization). Assume that the sampling frequency is , the sampling time is T, then the total number of data points is: ; The sampled signal data set is represented as: ;

[0052] Perform FFT transformation on the collected time domain vibration signal A(t) and convert it to the frequency domain to obtain the spectrum A(f): ; The amplitude spectrum calculated by FFT: ;in, is the kth frequency component, k=0,1,2,...,N / 2, They are Find the real and imaginary parts of the frequency component with the largest amplitude in the FFT result, that is, the main vibration frequency : ; Main vibration frequency Represents the main vibration mode of the construction structure.

[0053] Compare the changes in the main vibration frequency in adjacent time windows and calculate the vibration frequency change rate , the expression is: ;in: is the main vibration frequency of the i-th time window, is the main vibration frequency of the i-1th time window, Δt is the time window interval (for example, 5 seconds); calculate the vibration frequency change rate fluctuation index VFVFI, the expression is: ; Where: M is the total number of time windows in the calculation period, is the mean of the vibration frequency change rate in all windows.

[0054] The stability of workers' movement patterns is a key indicator of worker safety. Unlike traditional attendance or location information, this parameter focuses on the consistency and rationality of workers' movements, such as walking, standing, and climbing. It primarily analyzes movement characteristics using inertial sensors (IMUs) or depth cameras, and employs machine learning models (such as LSTM or HMM) for pattern recognition.

[0055] Under normal circumstances, the movement patterns of construction workers have certain regularities. For example, walking in the same work area is relatively stable, and climbing movements comply with operating specifications.

[0056] When the stability of the movement pattern decreases (for example, sudden discontinuous movement, abnormal stagnation, or fall signals occur), it may indicate that the person has slipped, lost balance, or even been affected by a sudden dangerous event, such as falling from a height or being hit by an object.

[0057] If the movement pattern becomes unstable for a long time, it may be due to fatigue, illegal operation (such as climbing without a safety belt), or a complex construction environment (such as slippery ground and increased obstacles). The system needs to further analyze its risk level.

[0058] The LSTM model is used to analyze the changes in the stability of the construction workers' movement patterns and generate a movement stability anomaly index. The movement stability anomaly index is obtained as follows:

[0059] Inertial sensors (IMU, including accelerometers, gyroscopes, and magnetometers) are used to sample the motion state of construction workers. The sensor data at each moment is represented as: ;in: is the triaxial acceleration (m / s²), is the three-axis angular velocity (° / s), is the three-axis magnetometer data (μT).

[0060] LSTM network training includes: input layer: input data is of shape (batch_size, N, 9), where 9 is the dimension of sensor data; LSTM layer: extracts time series features and outputs hidden states ; Fully connected layer: mapped to stability score (Between 0 and 1, indicating motion stability.) Loss function: The mean squared error (MSE) is used to calculate the error between the LSTM prediction and the true stability score to optimize the model.

[0061] Use the LSTM model to predict the stability score within each time window H Then, calculate the change rate of motion mode stability , the expression is: ;in: is the stability score of the current window, is the stability score of the previous window, Δt is the window interval (unit: seconds);

[0062] Calculate the Motion Stability Anomaly Index (MSAI). The Motion Stability Anomaly Index (MSAI) is used to measure the degree of change in a person's movement pattern and identify abnormal behavior. The calculation formula is: ; Where: Q is the total number of time windows, is the stability change rate of the i-th window, is the mean of the stability change rate in all windows.

[0063] The vibration frequency change rate fluctuation index and the motion stability anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each set of comprehensive feature vectors to predict the difference degree value label of the fused multi-source data as the prediction target, and takes minimizing the sum of the prediction errors of the difference degree value labels of all fused multi-source data as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The difference degree value of the fused multi-source data is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0064] The method for obtaining the difference degree value of the fused multi-source data is to obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model: Where, is the output function of the model, is the vibration frequency change rate fluctuation index, MSAI is the motion stability abnormality index, is the difference degree value of the fused multi-source data.

[0065] The obtained difference degree value of the fused multi-source data is compared with the difference degree reference threshold pre-set according to historical data. If the difference degree value of the fused multi-source data is greater than or equal to the pre-set difference degree reference threshold, it means that the difference degree of the fused multi-source data is high, and it is marked as multi-source data with high difference degree; if the difference degree value of the fused multi-source data is less than the pre-set difference degree reference threshold, it means that the difference degree of the fused multi-source data is low, and it is marked as multi-source data with low difference degree.

[0066] For multi-source data with high degree of difference, cluster analysis is used to further mine abnormal patterns. Specifically, the input data for cluster analysis comes from multi-source data samples with high degree of difference. Each sample contains multiple feature values, for example: ; Wherein: VFVFI is the vibration frequency change rate fluctuation index, MSAI is the motion stability abnormality index; is the main vibration frequency (Hz), is the rate of change of vibration frequency, is the standard deviation of the vibration change rate, Score the stability of LSTM predictions, is the rate of change of motion stability; all samples constitute a high-dimensional feature dataset: , J is the total number of features; Common clustering algorithms for finding abnormal patterns in high-variance data include: K-Means Clustering: Suitable for finding different types of abnormal patterns, such as equipment anomalies and personnel anomalies. DBSCAN (density-based spatial clustering): Suitable for mining rare but critical anomalies, such as extreme vibration or sudden falls. Gaussian Mixture Model (GMM): Suitable for handling situations where data may overlap, such as when abnormal human behavior may be related to environmental factors. K-Means clustering analysis can be used to divide high-variance data into multiple categories to identify key abnormal patterns.

[0067] Select the K value (you can use the Elbow Method or Silhouette Score).

[0068] Initialize K cluster centers and iteratively optimize: ;in, is the center of the k-th category, is the set of samples belonging to this category.

[0069] Calculate the Euclidean distance from each data point to the cluster center , and classify it into the nearest cluster: ; represents the jth eigenvalue of the i-th sample; Represents the value of the k-th cluster center on the j-th feature dimension; it is updated iteratively until convergence.

[0070] High-variance data at a construction site may be clustered into categories such as abnormal equipment vibration, unstable construction structure, and abnormal personnel behavior.

[0071] Each category can be further analyzed by the feature mean, for example: Category A (abnormal equipment vibration): high VFVFI, high ,Low Category B (abnormal personnel behavior): low VFVFI, high MSAI, high .

[0072] Based on cluster analysis results, intelligent early warning strategies can be developed. For high-risk categories (such as abnormal patterns discovered by K-Means), if a cluster persists for a long time, it indicates that the construction site may have systemic risks, such as aging equipment or illegal personnel operations. Measures such as equipment maintenance and construction code inspections can be implemented.

[0073] Identify security risks for multi-source data with low variance, and dynamically adjust early warning strategies at different levels based on risk levels and the variance of multi-source data, achieving real-time security warnings, including:

[0074] In the case of low variance, the sensor data and personnel movement data at the construction site are generally stable, but there may still be subtle risk signals. Therefore, it is necessary to conduct a fine-grained analysis of the following features:

[0075] A low-volume but continuously increasing rate of change (VFVFI) of vibration frequency may indicate long-term wear of equipment or structure rather than a sudden failure. A trend in the main vibration frequency deviation may indicate a gradual deterioration in the operating condition of construction equipment and the need for maintenance.

[0076] Minor abnormalities in movement patterns (e.g., a slight increase in MSAI but not reaching the high difference threshold) may indicate that construction workers are in a fatigued state and may cause safety accidents in the future.

[0077] Repeated violations (such as not wearing a helmet or short-term illegal operations) may not trigger a high-risk alarm at a single moment, but may increase the probability of accidents after accumulation.

[0078] Minor violations of air quality, temperature, humidity, and dust concentration may not immediately lead to safety incidents, but if they persist, they could have long-term impacts on worker health. Elevated noise levels, but not exceeding the danger threshold, could indicate aging construction equipment or subtle hidden dangers. By analyzing these subtle characteristics, the system can identify potential risk signals and assign a risk level.

[0079] Based on the results of the security risk assessment of low-difference degree data, the risks can be divided into low risks, medium risks and early warning level risks.

[0080] Low risk (safe state): The vibration frequency variation rate (VFVFI) is stable and there is no obvious increasing trend. The motion pattern anomaly index (MSAI) fluctuates little and there is no abnormal trend. The environmental parameters are within the safe range and there is no obvious fluctuation. The behavior patterns of construction workers are normal and there are no violation records. Disposal measures: The system continuously monitors, no intervention is required, only routine records are made.

[0081] Medium risk (needs attention state): The vibration frequency variation rate (VFVFI) has a slow upward trend, but still does not exceed the safety threshold. For example: 0.1 < VFVFI < 0.3 indicates that the construction equipment or structure may be in the early aging state.

[0082] The motion pattern anomaly index (MSAI) has slight fluctuations, but does not reach the level of falling or extreme behavior, for example: 0.1 < MSAI < 0.25; this may indicate that the work rhythm of construction workers is unstable. Construction workers have minor violations (such as not wearing safety helmets for a short time), but they do not occur continuously. Environmental factors are close to but do not exceed the safety threshold, for example, the dust concentration increases but does not reach the dangerous level. Disposal measures: Record potential risks and regularly review the data trend. If it is found that the trend continues to deteriorate, upgrade to the early warning level risk. Construction workers and equipment administrators receive reminders to pay attention to safe operation and equipment maintenance.

[0083] Early warning level risks (measures need to be taken) include:

[0084] The vibration frequency variation rate (VFVFI) exceeds the medium risk threshold and continues to rise: 0.3 ≤ VFVFI < 0.5 This may indicate that the equipment or structure has entered the stage of precursor to failure.

[0085] The motion pattern anomaly index (MSAI) continues to rise, but does not reach the level of sudden fall: 0.25 ≤ MSAI < 0.4; this may indicate that construction workers are in an over-fatigue state and safety accidents may occur in the future. Construction workers have repetitive violations, such as not wearing safety equipment many times, or showing short but frequent unstable behaviors (such as abnormal gait). Environmental parameters are close to dangerous values, such as noise level, dust concentration, etc. have approached the upper limit of the safety standard. Disposal measures: Trigger an automatic warning and notify the safety administrator. Require construction workers to take a short rest to reduce the probability of fatigue-related accidents. The equipment is maintained in advance to avoid small problems evolving into serious failures. If environmental factors continue to be close to dangerous values, adjust the construction arrangement to reduce the exposure risk.

[0086] Take the risk level and the difference degree value of multi-source data as the input items of fuzzy logic, and take early warning strategies at different levels as the output items of fuzzy logic, and perform dynamic regulation through fuzzy logic;

[0087] The fuzzy logic system needs to fuzzify the input items (risk level and the difference degree value CS), that is, define their membership functions for reasoning.

[0088] Input item 1: Risk level , and the risk level is defined as: Low (low risk): ≤0.3; Medium (medium risk): 0.3 < ≤0.7; High (high risk): > 0.7.

[0089] It is represented by a triangular membership function:

[0090] ;

[0091] ;

[0092] ;

[0093] The difference degree value CS of multi-source data, Low (low difference): CS ≤ 0.3; Medium (medium difference): 0.3 < CS ≤ 0.7; High (high difference): CS > 0.7; The definition of its membership function is similar to the expression.

[0094] Output early warning strategy: Mild Warning (mild early warning): Triggered when the risk is low and the difference is low. Moderate Warning (moderate early warning): Triggered when the risk is medium or the data has medium differences. Severe Warning (severe early warning): Triggered when the risk is high or the difference is high, and immediate action is required.

[0095] When the risk level is low (Low):

[0096] The data difference degree is low (Low): The construction status is stable and the risk is small. Therefore, only a mild early warning (Mild Warning) is required, and the system can run normally without additional intervention.

[0097] The data difference degree is medium (Medium): Although the data fluctuates slightly, it is still within the acceptable range, and a mild early warning (Mild Warning) can still be adopted to remind the management to pay attention to construction safety.

[0098] High: Although the risk level is low, the data discrepancy is high, which may indicate a potential problem. Therefore, a moderate warning is required. It is recommended to inspect the construction site to ensure there are no safety hazards.

[0099] When the risk level is medium:

[0100] Low data discrepancy: The overall construction status is relatively stable, but the risk level has increased and vigilance should be maintained. Therefore, a Mild Warning is used and regular inspections of the construction status are recommended.

[0101] Medium: Both the risk level and the data discrepancy are at a medium level, indicating that potential safety hazards may be accumulating. Therefore, a Moderate Warning should be used. On-site management personnel are advised to pay close attention and conduct equipment inspections and safety checks.

[0102] High Data Discrepancy: A high degree of data discrepancy, even at a medium risk level, could indicate an impending security incident. Therefore, a Severe Warning is issued, requiring immediate management intervention to investigate potential issues and develop safety precautions.

[0103] When the risk level is High:

[0104] Low data variance: Although the data has not changed much, the risk level is already high, so a moderate warning is still required to remind management personnel to remain vigilant and conduct inspections of the construction area.

[0105] Medium data discrepancy: At a high risk level, the data discrepancy is medium, indicating that the construction safety situation may be deteriorating. Therefore, a severe warning is required, and construction supervision must be strengthened. Temporary adjustments to the construction plan or equipment maintenance may be necessary.

[0106] High Data Discrepancy: When the risk level is high and the data discrepancy also reaches a high level, the probability of an accident is extremely high. Therefore, a Severe Warning must be issued immediately, which may require suspending construction, evacuating personnel, or implementing emergency measures to prevent a serious safety incident.

[0107] The Mamdani fuzzy reasoning method is used for dynamic adjustment to calculate the weights of different rules and obtain the final warning level.

[0108] The final warning level value is calculated using the Centroid Method: ;in: is the membership value corresponding to different rules, It is a numerical mapping of different warning strategies (for example, Mild=1, Moderate=2, Severe=3).

[0109] Based on the fuzzy inference results, the system dynamically implements corresponding warning measures: Mild Warning: Only a prompt message is sent to alert management personnel. Moderate Warning: Equipment inspections and personnel safety checks are recommended. Severe Warning: Immediately notify management personnel, suspend construction, and initiate safety emergency plans.

[0110] The terminal interaction module is the core output interface of the intelligent construction site safety inspection system. It is responsible for promptly pushing safety inspection results and warning information to management terminals, enabling remote monitoring and efficient decision-making. This module integrates a mobile app, a monitoring center display, and voice broadcast equipment, ensuring that managers can monitor the safety status of the construction site anytime, anywhere and take appropriate safety management measures. By synchronizing information across multiple terminals, managers can obtain real-time information on safety inspection results, warning levels, risk assessment data, and historical records for the construction area, thereby improving the scientific nature and accuracy of construction safety management.

[0111] On the mobile app side, the terminal interaction module supports remote access to real-time monitoring data from the construction site, including video surveillance footage, safety inspection reports, equipment operating status, and personnel behavior analysis. Managers can use the app to receive real-time alerts, view detailed data on various safety incidents, and remotely confirm or assign tasks. The system also supports voice or text notifications, allowing managers to obtain critical safety information immediately, even when away from the monitoring center, enabling rapid response to emergencies.

[0112] Regarding the monitoring center's display screens and voice announcements, the terminal interaction module visualizes the construction site's safety status on a large display screen, including safety status distribution maps, risk trend analysis, and alarm event lists. This allows on-duty personnel to intuitively review the safety situation across the entire construction area. The system also supports voice announcements. When high-risk events (such as falls from height, equipment failures, or fire hazards) are detected, an automatic voice alarm is triggered, prompting on-site personnel and management to take prompt action. Through these interactive methods, the terminal interaction module significantly improves the efficiency of construction site safety supervision and reduces the probability of safety accidents.

[0113] This embodiment includes four modules: data acquisition, data fusion and processing, intelligent analysis and early warning, and terminal interaction, to achieve intelligent safety inspections at construction sites. The data acquisition module is responsible for acquiring multi-source data from the construction site, including video surveillance, sensor data (temperature, humidity, dust concentration, vibration intensity), RFID identification information, and environmental data. The data fusion and processing module synchronizes the time, standardizes the format, and filters noise from the collected data, employing an adaptive weighted fusion algorithm to optimize multi-source data fusion. The intelligent analysis and early warning module uses machine learning to assess data variance. For high-variance data, cluster analysis is used to identify abnormal patterns. For low-variance data, safety risks are identified and early warning strategies are dynamically adjusted based on risk levels and data changes, achieving real-time safety warnings. The terminal interaction module pushes inspection results and early warning information to mobile apps, monitoring center displays, and voice broadcast devices. Managers can remotely monitor construction status and make safety decisions, improving safety oversight capabilities at construction sites.

[0114] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0115] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0116] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. An intelligent safety inspection system for construction sites, characterized by: It includes data acquisition module, data fusion and processing module, intelligent analysis and early warning module and terminal interaction module; A data acquisition module is used to acquire multi-source data from the construction site, including video surveillance data, sensor data, RFID identification information, and environmental data. The sensor data includes temperature, humidity, dust concentration, and vibration intensity data. The data fusion and processing module is used to synchronize the time of the collected multi-source data, standardize the data format, and filter noise, and dynamically optimize and fuse different data sources based on an adaptive weighted fusion algorithm; The intelligent analysis and early warning module is used to evaluate the degree of difference of fused multi-source data based on machine learning models. For multi-source data with high differences, cluster analysis is used to further explore abnormal patterns. For multi-source data with low differences, security risks are identified and early warning strategies of different levels are dynamically adjusted based on the risk level and the changes in the degree of difference of multi-source data, achieving real-time security warnings. Specifically, the method includes: converting the vibration frequency change rate fluctuation index and the motion stability anomaly index into a comprehensive feature vector, using the comprehensive feature vector as the input of a machine learning model, using each set of comprehensive feature vectors to predict the difference degree value label of the fused multi-source data as a prediction target, minimizing the sum of the prediction errors of the difference degree value labels of all fused multi-source data as a training target, training the machine learning model until the sum of the prediction errors reaches convergence, stopping the model training, and determining the difference degree value of the fused multi-source data based on the model output results, wherein the machine learning model is a polynomial regression model; Compare the obtained difference degree value of the fused multi-source data with the difference degree reference threshold pre-set based on historical data. If the difference degree value of the fused multi-source data is greater than or equal to the pre-set difference degree reference threshold, it indicates that the difference degree of the fused multi-source data is high, and it is marked as multi-source data with high difference degree; if the difference degree value of the fused multi-source data is less than the pre-set difference degree reference threshold, it indicates that the difference degree of the fused multi-source data is low, and it is marked as multi-source data with low difference degree; Among them, the LSTM model is used to analyze the changes in the stability of the construction workers' movement patterns to generate a movement stability anomaly index. The method for obtaining the movement stability anomaly index is as follows: The inertial sensor is used to sample the motion state of the construction workers. The sensor data at each moment is expressed as: ;in: is the triaxial acceleration, is the three-axis angular velocity, is the three-axis magnetometer data; LSTM network training includes: input layer: input data is shape; LSTM layer: extract time series features and output hidden state ; Fully connected layer: mapped to stability score ; Loss function: Use mean square error to calculate the error between the LSTM prediction value and the true stability score to optimize the model; Use the LSTM model to predict the stability score within each time window H Then, calculate the change rate of motion mode stability , the expression is: ;in: is the stability score of the current window, is the stability score of the previous window, Δt is the window interval; the motion stability abnormality index MSAI is calculated using the following formula: ; Where: Q is the total number of time windows, is the stability change rate of the i-th window, is the mean of the stability change rate in all windows; The terminal interaction module is used to push safety inspection results and warning information to the management terminal, including mobile APP, monitoring center display screen and voice broadcast equipment. Management personnel can remotely view the construction site conditions and make safety decisions through the terminal interaction module.

2. The intelligent safety inspection system for construction sites according to claim 1, characterized in that: In the intelligent analysis and early warning module, the vibration frequency change rate is analyzed by fast Fourier transform to generate the vibration frequency change rate fluctuation index. The method for obtaining the vibration frequency change rate fluctuation index is as follows: A high-precision acceleration sensor is used to collect the vibration signal of the construction structure in time series. The collected time domain vibration signal A(t) is transformed by FFT and converted into the frequency domain to obtain the spectrum A(f): ; The amplitude spectrum calculated by FFT: ;in, is the kth frequency component, k=0,1,2,...,N / 2, They are Find the real and imaginary parts of the frequency component with the largest amplitude in the FFT result, that is, the main vibration frequency : ; Compare the changes in the main vibration frequency in adjacent time windows and calculate the vibration frequency change rate , the expression is: ;in: is the main vibration frequency of the i-th time window, For the The main vibration frequency of the time window is Δt, and the time window interval is Δt. The vibration frequency change rate fluctuation index VFVFI is calculated, and the expression is: ; Where: M is the total number of time windows in the calculation period, is the mean of the vibration frequency change rate in all windows.

3. The intelligent safety inspection system for construction sites according to claim 1, characterized in that: For multi-source data with high degree of variation, cluster analysis is used to further mine abnormal patterns, specifically: The input data for cluster analysis comes from multi-source data samples with high degree of difference, and each sample contains multiple feature values: ; Wherein: VFVFI is the vibration frequency change rate fluctuation index, MSAI is the motion stability abnormality index; The main vibration frequency, is the rate of change of vibration frequency, is the standard deviation of the vibration change rate, Score the stability of LSTM predictions, is the rate of change of motion stability; all samples constitute a high-dimensional feature dataset: , J is the total number of features; Select the K value, initialize K cluster centers, and iteratively optimize: ;in, is the center of the k-th category, is the set of samples belonging to this category; Calculate the Euclidean distance from each data point to the cluster center , and classify it into the nearest cluster: ; represents the jth eigenvalue of the i-th sample; Represents the value of the k-th cluster center on the j-th feature dimension; iteratively update until convergence; High-variance data from construction sites are clustered into categories such as abnormal equipment vibration, unstable construction structure, and abnormal personnel behavior.

4. The intelligent safety inspection system for construction sites according to claim 3, characterized in that: Dynamically adjust early warning strategies at different levels based on risk levels and the degree of difference in multi-source data to achieve real-time security early warnings. Specifically: The risk level and the difference between multi-source data are used as the input of fuzzy logic, and the early warning strategies of different levels are used as the output of fuzzy logic, and dynamic regulation is carried out through fuzzy logic. The fuzzy membership function is used to fuzzify the input, the fuzzy rule base is used for reasoning, and the center of gravity method is used to defuzzify and calculate the final warning level value. , ;in: is the membership value corresponding to different rules, Numerical mapping for different early warning strategies; According to the results of fuzzy reasoning, corresponding early warning measures are dynamically executed.

Images