Highway construction safety monitoring system and method

Abstract

The invention discloses a highway construction safety monitoring system and method, and belongs to the technical field of highway construction. The method comprises the following steps of 1, collecting construction site data in real time, wherein the construction site data comprises environment data, video data and construction management data; step 2, generating global situation awareness data; 3, evaluating the safety risk probability and severity in real time, and predicting a risk propagation path and a cascade effect; step 4, generating a plurality of risk disposal schemes for sorting and screening, and selecting an optimal disposal suggestion; 5, precise early warning and differentiated management are achieved, the false alarm rate is reduced, and high-risk events are found in time; and step 6, displaying the construction site through three-dimensional reconstruction, superposing construction site data and a risk assessment result, and providing remote monitoring and emergency command of a mobile terminal.

Description

Technical Field

[0001] The present application relates to the technical field of highway construction, and more specifically, to a highway construction safety monitoring system and method. Background Art

[0002] Highway construction safety has always been an important part of construction management, which is related to the life safety of construction workers and the quality of the project. However, as the scale of modern highway construction projects continues to expand, the complexity and uncertainty of the construction environment are also increasing, and traditional safety monitoring methods have been unable to cope with new challenges.

[0003] At present, the safety monitoring of highway construction sites mainly relies on manual inspections and fixed monitoring equipment. These traditional methods not only have poor real-time performance and limited coverage, but also often have omissions or false alarms due to human factors. In addition, with the development of technologies such as the Internet of Things (IoT), big data, artificial intelligence (AI), and 3D reconstruction, the amount of data on construction sites has increased exponentially, and traditional methods are difficult to effectively process and analyze this data.

[0004] In recent years, intelligent construction safety monitoring technology has gradually been applied. By deploying various sensors and cameras at the construction site, real-time environmental data, video data, and construction management data are collected, and edge computing technology is used for preliminary processing and transmission, providing a new idea for construction safety monitoring. However, the existing intelligent monitoring system still has many shortcomings in multi-source data fusion, real-time risk assessment, and emergency response, and it is difficult to meet the safety needs in complex construction environments.

[0005] In this context, there is an urgent need for a highway construction safety monitoring system that can comprehensively use technologies such as the Internet of Things, artificial intelligence, big data, and three-dimensional reconstruction to achieve real-time collection, fusion, and analysis of multi-source data, conduct all-round and real-time monitoring of construction sites, and provide accurate risk assessments and emergency response plans.

[0006] To sum up, how to realize real-time collection and analysis of multi-source data, accurate risk assessment and efficient emergency response in highway construction has become a technical problem that needs to be solved urgently. Summary of the invention

[0007] In order to overcome a series of defects in the prior art, the purpose of this application is to provide a highway construction safety monitoring method for the above-mentioned problems, including the following steps.

[0008] Step 1: Collect construction site data in real time, including environmental data, video data, and construction management data.

[0009] Step 2: Generate global situational awareness data.

[0010] Step 3: Assess the probability and severity of safety risks in real time, and predict the risk transmission path and cascading effects.

[0011] Step 4: Generate multiple risk treatment plans, sort and screen them, and select the best treatment suggestion.

[0012] Step 5: Achieve accurate early warning and differentiated management, reduce false alarm rates and detect high-risk events in a timely manner.

[0013] Step 6: Display the construction site through 3D reconstruction, overlay the construction site data and risk assessment results, and provide mobile terminal remote monitoring and emergency command.

[0014] Furthermore, step 1 includes the following steps.

[0015] Multiple environmental monitoring units, including temperature, humidity, noise and air quality sensors, are deployed at the construction site to establish a real-time data collection network.

[0016] High-definition cameras and intelligent monitoring equipment are installed at key locations on the construction site to record the construction process in real time, including workers' operations, mechanical equipment operating status, and material stacking and usage.

[0017] Capture construction progress, staffing, equipment usage, and material consumption data from construction management systems.

[0018] All kinds of data are transmitted to the data processing center in real time via wireless networks.

[0019] The collected raw data is preliminarily screened, cleaned and classified to eliminate abnormal data.

[0020] Establish a data backup mechanism and regularly back up all collected data.

[0021] Furthermore, step 2 includes the following steps.

[0022] The collected environmental monitoring data, video data and construction management data are time-stamped and synchronized to ensure that all data are aligned in the time dimension and establish a unified time reference.

[0023] By accurately locating monitoring points, camera locations and construction areas, we can achieve spatial correlation of data and build a complete spatiotemporal data system.

[0024] Use machine vision and deep learning algorithms to extract features from video data and identify key objects, events, and behavior patterns at the construction site, including dynamic tracking of construction personnel and machinery, and automatic detection and marking of abnormal situations.

[0025] Conduct multi-dimensional feature analysis on environmental monitoring data, extract temperature change trends, noise fluctuation patterns and air quality index characteristics, establish a correlation model between environmental factors, and quantitatively describe the environmental situation.

[0026] The extracted features are deeply integrated to build a multi-source information fusion model to generate standardized situational awareness data including personnel safety, equipment status, environmental conditions and construction progress.

[0027] Furthermore, the extracted features are deeply integrated to construct a multi-source information fusion model to generate standardized situational awareness data including personnel safety, equipment status, environmental conditions, and construction progress, including the following steps.

[0028] Use computer vision algorithms to detect and track targets in construction site videos, identify the location and dynamic information of construction personnel, equipment, and materials, perform semantic segmentation on construction scenes, and store the extracted feature data in a unified standard format.

[0029] The environmental characteristic data is cleaned and standardized, time series features are constructed, and the environmental characteristic map of the construction site is generated in combination with the sensor location information.

[0030] Structural processing of construction management data is carried out, unified data standards are established, and unstructured text information is converted into a quantifiable indicator system.

[0031] A unified indexing system is established based on timestamps and spatial coordinates to align video features, environmental features, and management data in time and space dimensions.

[0032] The aligned multi-source feature data are input into the deep neural network model, and a unified feature representation is generated through multi-level feature extraction and fusion, and the corresponding output layer is designed according to the specific needs of the construction site.

[0033] The output results of the deep neural network model are post-processed to form a comprehensive situational awareness data set that includes personnel safety assessment, equipment operating status, environmental condition evaluation, and construction progress quantitative indicators. A visual display interface is established to facilitate real-time monitoring and decision-making by managers.

[0034] Furthermore, step 3 includes the following steps.

[0035] Conduct in-depth mining of environmental monitoring data and safety accident records from past construction projects, identify the correlation between environmental factors and the occurrence of safety accidents, and establish an environment-accident correlation model.

[0036] Compare the currently collected environmental data with the historical data model in real time, calculate the probability of safety risks occurring and generate a standardized risk score.

[0037] Based on situational awareness data, potential risks are quantitatively graded from the dimensions of personal injury, economic loss and construction delay.

[0038] Predict the risk diffusion path and impact scope, and simulate the evolution of risks in time and space dimensions.

[0039] Assess the possibility and impact of chain reactions and secondary disasters, and identify key nodes and unstable links.

[0040] Furthermore, step 4 includes the following steps.

[0041] Establish a knowledge base of risk management solutions, organize the experience and best practices in handling various risk events in historical projects, and extract standard handling processes and key elements for different risk types.

[0042] Based on the current risk assessment results and predicted risk propagation paths, similar cases are quickly retrieved from the knowledge base, and multiple feasible disposal plans are automatically generated based on the actual situation on site.

[0043] Multi-dimensional evaluation indicators are set, including implementation cost, time efficiency, resource consumption, and safety and reliability, to quantitatively score and comprehensively rank each candidate solution.

[0044] Conduct virtual drills and effectiveness evaluations on the top-ranked disposal plans to predict the implementation effects of each plan under different conditions and potential problems.

[0045] Based on the comprehensive evaluation results, the optimal disposal plan is screened out and a detailed implementation proposal is formed, which includes specific operating steps, required resources, personnel division of labor and emergency plan.

[0046] Furthermore, step 5 includes the following steps.

[0047] Based on historical data analysis and current risk assessment results, a dynamic adjustment mechanism for warning thresholds is established, including risk level classification standards and warning trigger conditions.

[0048] Develop differentiated management strategies based on construction stages, formulate corresponding early warning standards and management requirements according to the characteristics of each stage, and automatically adjust management priorities according to the current construction progress to achieve accurate risk prevention and control.

[0049] Build a weather-sensitive early warning mechanism, access meteorological forecast data, predict the impact of weather changes on construction safety in advance, and automatically adjust the warning level and response measures according to different weather conditions.

[0050] Continuously analyze the correlation between warning records and actual accidents to identify and eliminate false warning features.

[0051] Conduct real-time analysis and verification of abnormal data that exceeds the normal fluctuation range to ensure that major risks and hidden dangers are discovered and dealt with in a timely manner.

[0052] Furthermore, step 6 includes the following steps.

[0053] Based on multi-source data collection, photogrammetry and laser scanning technology are combined to perform high-precision three-dimensional reconstruction of the construction site, build a digital twin model that includes terrain, facilities and equipment elements, and update the data regularly to ensure dynamic synchronization with the site, providing a spatial reference for risk management.

[0054] The real-time collected environmental data, construction management data and risk assessment results are integrated into the digital twin model in the form of layer overlays, and the spatial distribution and change trends of various types of data are intuitively displayed through color coding or symbol annotation.

[0055] The locations of identified risk sources are marked in real time in the digital twin model, and the risk level and impact scope are displayed to achieve accurate risk management and control.

[0056] Through the playback and analysis of spatiotemporal data, we conduct in-depth research on the occurrence and development patterns of risk events, summarize experiences and lessons, and optimize risk prevention and control strategies.

[0057] The three-dimensional scene and risk monitoring functions are extended to mobile terminals, supporting managers to remotely view on-site conditions, receive risk warnings and issue emergency instructions.

[0058] The purpose of this application is also to provide a highway construction safety monitoring system, including the following modules.

[0059] The data acquisition module collects environmental data, video data and construction management data from the construction site in real time, and performs preliminary processing and transmission through edge computing.

[0060] The data processing module performs spatiotemporal alignment, feature extraction and standardization on multi-source heterogeneous data, builds a unified spatiotemporal data system, and realizes deep data fusion and feature analysis.

[0061] The risk assessment module, based on historical data mining and deep learning algorithms, analyzes the correlation between environmental factors and safety accidents, evaluates risk probability and severity in real time, and builds a risk propagation model.

[0062] The solution generation module retrieves and generates multiple risk disposal solutions from the knowledge base based on risk assessment results and historical experience, and optimizes sorting and screening through multi-dimensional evaluation to form an executable detailed implementation plan.

[0063] The execution monitoring module is responsible for the specific implementation of the risk disposal plan, including tracking the progress of plan execution, evaluating the execution effect and providing problem feedback, and providing real-time feedback of the execution status to relevant modules.

[0064] The early warning decision-making module dynamically formulates and adjusts early warning rules and thresholds based on seasonal, weather and construction stage factors and combined with risk assessment results to achieve hierarchical early warning and differentiated management strategies.

[0065] The intelligent interactive module visualizes construction site data and risk assessment results through 3D reconstruction and digital twin technology, and supports risk source location, historical backtracking, and remote monitoring and command on mobile terminals.

[0066] Compared with the prior art, the present application has the following beneficial effects.

[0067] This application generates global situational awareness data by real-time collection and integration of multi-source data, real-time assessment and prediction of safety risks, generation and screening of multiple risk management plans, and accurate early warning and differentiated management. It also uses three-dimensional reconstruction technology to display the construction site and provide mobile remote monitoring and emergency command, thereby comprehensively improving the accuracy and efficiency of construction safety monitoring, reducing the false alarm rate, and strengthening the timely detection and handling of high-risk events. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic flow chart of a highway construction safety monitoring method disclosed in an embodiment of the present application.

[0069] Figure 2 This is a flow chart of step 3 in the embodiment of the present application.

[0070] Figure 3 This is a schematic diagram of the structure of a highway construction safety monitoring system disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical scheme and advantages of the implementation of the present invention clearer, the technical scheme in the embodiment of the present invention will be described in more detail below in conjunction with the drawings in the embodiment of the present invention. In the drawings, the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The described embodiments are part of the embodiments of the present invention, not all of the embodiments.

[0072] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0073] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0074] like Figure 1 As shown, a highway construction safety monitoring method includes the following steps.

[0075] Step 1: Collect construction site data in real time, including environmental data, video data, and construction management data.

[0076] Step 2: Perform spatiotemporal alignment and feature extraction on the collected environmental data, video data, and construction management data to generate global situation awareness data.

[0077] Step 3: Based on historical data, learn the relationship between environmental factors and safety accidents, evaluate the probability and severity of safety risks in real time, and predict the risk propagation path and cascade effect.

[0078] Step 4: Based on the risk assessment results and historical experience, generate multiple risk treatment plans for sorting and screening, and select the best treatment suggestion.

[0079] Step 5: Consider seasonal, weather and construction stage factors, dynamically adjust the warning threshold, achieve accurate warning and differentiated management, reduce false alarm rates and detect high-risk events in a timely manner.

[0080] Step 6: Display the construction site through 3D reconstruction, overlay the construction site data and risk assessment results, support risk source location, historical backtracking and decision simulation, and provide mobile remote monitoring and emergency command.

[0081] The highway construction safety monitoring method disclosed in this embodiment greatly improves the intelligence and accuracy of construction site safety management through multi-step comprehensive analysis. First, various types of construction data (environment, video and construction management data) are collected in real time, providing comprehensive basic information for subsequent analysis. In the data processing link, global situation awareness data is constructed through spatiotemporal alignment and feature extraction, providing rich background information for decision-making. Through historical data analysis, the probability of safety risks and the path of risk propagation can be evaluated in real time, the early warning capability can be strengthened, and potential high-risk events can be effectively identified. In addition, based on comprehensive risk assessment, a variety of risk disposal plans can be generated, and the optimal disposal plan can be recommended through the optimization algorithm, which is helpful to respond to emergencies in a timely manner. The dynamic adjustment of the warning threshold function can optimize the risk warning strategy according to environmental changes, reduce the false alarm rate, and ensure the sensitivity detection of high-risk events. Finally, the addition of three-dimensional reconstruction and remote monitoring functions enables decision makers to understand the risk status of the construction site more intuitively, improves the ability to locate risk sources, historical backtracking and emergency response, and further enhances the real-time and emergency command efficiency. Overall, this method achieves accurate management of construction safety and efficient emergency handling through multi-data fusion, intelligent analysis and dynamic adjustment.

[0082] Furthermore, step 1 includes the following steps.

[0083] Multiple environmental monitoring units, including temperature, humidity, noise and air quality sensors, are deployed at the construction site to establish a real-time data collection network to ensure the continuity and accuracy of environmental data.

[0084] High-definition cameras and intelligent monitoring equipment are installed at key locations on the construction site to record the construction process in real time, including workers' operations, mechanical equipment operating status, and material stacking and usage.

[0085] Capture construction progress, staffing, equipment usage, and material consumption data from construction management systems.

[0086] All kinds of data are transmitted to the data processing center in real time via wireless networks to ensure the stability and reliability of data transmission.

[0087] The collected raw data is preliminarily screened, cleaned and classified to eliminate abnormal data, laying the foundation for subsequent analysis and application.

[0088] Establish a data backup mechanism and regularly back up all collected data to ensure data security and facilitate subsequent query and analysis.

[0089] The above steps ensure the comprehensiveness, accuracy and efficient transmission of construction site data through a variety of data collection and processing methods. First, by deploying a variety of environmental monitoring units at the construction site, a real-time data collection network is established to ensure the continuity and accuracy of environmental data. By installing high-definition cameras and intelligent monitoring equipment, the key links of the construction process (such as worker operations, equipment operation and material use) are fully recorded, enhancing the comprehensive monitoring of the site conditions. Combined with the management data such as construction progress, staffing, and equipment use obtained by the construction management system, the data source is further enriched, providing a multi-dimensional understanding of the on-site construction situation. All data are transmitted to the data processing center in real time via a wireless network, ensuring the stability and reliability of data transmission and the feasibility of real-time monitoring and analysis. After receiving the original data, the data processing center conducts preliminary screening, cleaning and classification to eliminate abnormal data, thereby ensuring the accuracy of subsequent analysis. At the same time, the establishment of a data backup mechanism also provides a guarantee for the safe storage of data, ensuring the long-term traceability of construction data, and facilitating future inquiries and in-depth analysis. Overall, the above steps provide high-quality basic data support for intelligent monitoring and risk assessment while improving data collection and processing efficiency and ensuring data accuracy and security.

[0090] Specifically, an environmental monitoring unit is set up for every 500 square meters of construction area. Each environmental monitoring unit contains: 2 temperature sensors with a sampling frequency of 300 seconds / time, a measurement range of -40℃ to 85℃, and an accuracy of ±0.5℃; 2 humidity sensors with a sampling frequency of 300 seconds / time, a measurement range of 0-100%RH, and an accuracy of ±3%RH; 4 noise sensors with a sampling frequency of 10 seconds / time, a measurement range of 30-120dB, and an accuracy of ±2dB; 2 air quality sensors with a sampling frequency of 600 seconds / time, which can monitor PM2.5 concentrations in the range of 0-500μg / m³, an accuracy of ±15μg / m³, and CO2 concentrations in the range of 0-5000ppm, with an accuracy of ±50ppm. It can be seen that the environmental monitoring unit can comprehensively monitor the environmental conditions of the construction site through the high-frequency sampling and accuracy of multiple sensors, and provide key data support for safety warnings and risk assessments. In addition, accurate data collection and monitoring can effectively provide a technical basis for construction management and personnel safety, ensuring that environmental safety and health standards are met during the construction process.

[0091] Specifically, the deployment and technical parameters of high-definition cameras include: no less than 4 monitoring points are set for every 1,000 square meters of construction area; video resolution supports 1920×1080P and 2560×1440P adjustable; frame rate supports 15fps and 25fps adjustable; horizontal field of view angle is 90 degrees, vertical field of view angle is 50 degrees; support 350-degree horizontal rotation and -5 degrees to 85 degrees pitch adjustment; infrared night vision distance is not less than 30 meters; protection level meets IP65 standard; support H.264 and H.265 video encoding standards. Overall, this technical configuration provides a high-definition, all-round and stable monitoring system for the construction site, which can monitor the construction site in real time, capture key events, ensure construction safety, and provide important video data support for later risk assessment and decision-making.

[0092] Furthermore, step 2 includes the following steps.

[0093] The collected environmental monitoring data, video data and construction management data are time-stamped and synchronized to ensure that all data are aligned in the time dimension and establish a unified time reference.

[0094] By accurately locating monitoring points, camera locations and construction areas, we can achieve spatial correlation of data and build a complete spatiotemporal data system.

[0095] Use machine vision and deep learning algorithms to extract features from video data and identify key objects, events, and behavior patterns at the construction site, including dynamic tracking of construction personnel and machinery, and automatic detection and marking of abnormal situations.

[0096] Conduct multi-dimensional feature analysis on environmental monitoring data, extract temperature change trends, noise fluctuation patterns and air quality index characteristics, establish a correlation model between environmental factors, and quantitatively describe the environmental situation.

[0097] The extracted features are deeply integrated to build a multi-source information fusion model to generate standardized situational awareness data including personnel safety, equipment status, environmental conditions and construction progress.

[0098] The above steps greatly improve the integration and intelligent analysis capabilities of construction site data through precise time and space synchronization processing, as well as multi-dimensional data analysis and fusion. First, through timestamp calibration and synchronization processing, the consistency of environmental monitoring data, video data and construction management data in the time dimension is ensured, providing a unified time benchmark for subsequent analysis and decision-making, and avoiding data deviations caused by timing problems. Secondly, by accurately locating monitoring points, camera positions and construction areas, the spatial association of data is realized, and a spatiotemporal data system is constructed, so that various types of data can be effectively combined in space, thereby fully reflecting the dynamic changes of the construction site. In terms of video data processing, feature extraction using machine vision and deep learning algorithms can accurately identify key objects, events and behavior patterns at the construction site, such as dynamic tracking of construction personnel and mechanical equipment, as well as automatic detection and marking of abnormal situations. This automated video analysis can efficiently identify potential safety hazards, reduce the workload of manual monitoring, and improve safety management efficiency. For environmental monitoring data, through multi-dimensional feature analysis, temperature change trends, noise fluctuation patterns and air quality index characteristics can be extracted, and a correlation model between environmental factors can be established, providing data support for the quantitative description of environmental trends. Finally, the extracted video features, environmental features and construction management data were deeply integrated to build a multi-source information fusion model and generate standardized situational awareness data. This integrated model not only includes multi-dimensional information such as personnel safety, equipment status, environmental conditions and construction progress, but also provides a comprehensive and accurate data foundation for real-time monitoring, decision support and risk assessment. Overall, step 2 significantly improved the safety, efficiency and management level of the construction site through multi-source data fusion, intelligent analysis and accurate modeling.

[0099] Furthermore, by accurately locating monitoring points, camera locations, and construction areas, spatial correlation of data is achieved and a complete spatiotemporal data system is constructed, including the following steps.

[0100] Establish a global coordinate system for the construction area to describe the set of monitoring points {P i}、Camera collection {C j} and the key point set of the construction area {R k}’s spatial position relationship, ensuring that all spatial data are analyzed in the same reference system.

[0101] Based on the global coordinate system, calculate the Euclidean distance d(P i ,C j ) and the Euclidean distance d(C j ,R k ), and according to the preset threshold D max Conduct spatial correlation screening and establish preliminary spatial topological relationships.

[0102] Add a timestamp t to each data point to form a four-dimensional space-time point P i (t), C j (t) and R k (t), and by calculating the time difference |t i -t j ∣Filter out data pairs with time series association to ensure the real-time nature of data association.

[0103] Taking into account the time difference and spatial distance, based on the formula Similarity(P i ,C j ,R k )=1 / [1+α·|t i -t j |+β·d(P i ,C j )+γ·d(C j ,R k )] Calculate the similarity between monitoring points, cameras and key points in the construction area, and select point pairs with high similarity for precise association, where: Similarity(P i ,C j ,R k ) is the monitoring point P i 、Camera C j and construction area key points R k is the similarity measure between them, indicating their degree of association in space and time; α is the weight coefficient of time difference, which controls the influence of time difference on similarity calculation; t i is the timestamp of the i-th monitoring point, indicating the time when the monitoring point collected data; t j is the timestamp of the jth camera, indicating the time when the camera collects data; β is the spatial distance d(P i ,C j ) is the weight coefficient that controls the influence of the distance between the monitoring point and the camera on the similarity; γ is the spatial distance d(C j ,R k ) to control the influence of the distance between the camera and the construction area on the similarity.

[0104] The filtered high-relevance data are fused to build a global visualization view of the real-time status of the construction site.

[0105] In this embodiment, a complete spatiotemporal data system is successfully constructed through precise spatial positioning and temporal association, and efficient association and fusion of construction site data are achieved. First, by establishing a global coordinate system, the key points of the monitoring points, cameras, and construction areas are accurately mapped to a unified spatial reference system, laying the foundation for the association analysis of spatial data. Then, based on the global coordinate system, the Euclidean distance between the monitoring point and the camera, and between the camera and the key points of the construction area are calculated, and the spatial topological relationship is initially established to ensure the effective connection of the data in space. In terms of temporal association, a timestamp is added to each data point, and the data pairs with temporal correlation are screened out by calculating the time difference, which ensures the real-time nature of data association and avoids data lag or mismatch. By comprehensively considering the time difference and spatial distance, the similarity measurement formula is used to further accurately calculate the correlation between the monitoring point, camera, and key points of the construction area, and optimize the association process of the data points. The weight coefficients (α, β, γ) in the formula enable the influence of time difference and spatial distance on similarity to be flexibly adjusted, thereby ensuring that the point pairs with high correlation can be accurately selected and associated. Finally, by fusing the screened high-relevance data, a global visualization of the real-time status of the construction site can be constructed, allowing decision makers to intuitively understand the safety situation, progress and potential risks of the site.

[0106] Furthermore, the extracted features are deeply integrated to construct a multi-source information fusion model to generate standardized situational awareness data including personnel safety, equipment status, environmental conditions, and construction progress, including the following steps.

[0107] Use computer vision algorithms to detect and track targets in construction site videos, identify the location and dynamic information of construction personnel, equipment, and materials, perform semantic segmentation on construction scenes, and store the extracted feature data in a unified standard format.

[0108] The environmental characteristic data is cleaned and standardized, time series features are constructed, and the environmental characteristic map of the construction site is generated in combination with the sensor location information.

[0109] Structural processing of construction management data is carried out, unified data standards are established, and unstructured text information is converted into a quantifiable indicator system.

[0110] A unified indexing system is established based on timestamps and spatial coordinates to align video features, environmental features, and management data in time and space dimensions to ensure consistency of data from different sources.

[0111] The aligned multi-source feature data are input into the deep neural network model, and a unified feature representation is generated through multi-level feature extraction and fusion, and the corresponding output layer is designed according to the specific needs of the construction site.

[0112] The output results of the deep neural network model are post-processed to form a comprehensive situational awareness data set that includes personnel safety assessment, equipment operating status, environmental condition evaluation, and construction progress quantitative indicators. A visual display interface is established to facilitate real-time monitoring and decision-making by managers.

[0113] In this embodiment, through the deep integration and intelligent analysis of multi-source information, the intelligent level of construction site safety management, equipment monitoring and environmental assessment is significantly improved. First, the video data of the construction site is detected and tracked by computer vision algorithms, which can accurately identify the real-time location and dynamic changes of construction personnel, equipment and materials, and provide key visual data for real-time monitoring. In addition, the semantic segmentation technology of the construction scene extracts and stores different scene information in a structured manner to ensure that the feature data can be processed and used according to unified standards. In terms of environmental feature processing, an environmental feature map of the construction site is generated by cleaning and standardizing sensor data, and the spatial expression of environmental data is further enhanced in combination with the sensor position. After structured processing, the construction management data converts unstructured text information into a quantifiable indicator system, so that various types of management data can be seamlessly combined with other information sources, enhancing the comparability and consistency of the data. The unified index system established based on timestamps and spatial coordinates further ensures the alignment of each data source in time and space dimensions, laying the foundation for subsequent data fusion. Then, the deep neural network model generates a unified feature representation through multi-level feature extraction and fusion, so that multi-source data can be efficiently integrated, further improving the accuracy and depth of situational awareness. After post-processing, the output results of the model form a comprehensive situational awareness data set including personnel safety assessment, equipment status monitoring, environmental condition assessment and construction progress, which greatly enhances the ability of real-time monitoring of the construction site. At the same time, the equipped visual display interface enables managers to intuitively view real-time data, make decisions and respond quickly, and improve the management efficiency and safety of the construction site. Overall, through deep learning and multi-source data fusion, it provides comprehensive and accurate intelligent support for safety management and operation optimization of the construction site.

[0114] Specifically, the deep neural network model adopts: 8-layer convolutional neural network structure; the convolution layer uses 3×3 and 5×5 convolution kernels alternately set; 2×2 maximum pooling layer and batch normalization layer; the number of neurons in the fully connected layer is 2048, 1024, and 512 respectively; LeakyReLU activation function is used; the dropout rate is set to 0.3; the ratio of training set, validation set, and test set is 6:2:2; the basic training samples are not less than 50,000; the initial value of the learning rate is set to 0.001, and the cosine annealing strategy is used; the batch size is set to 64. It can be seen that the deep neural network model effectively improves the accuracy and robustness of complex construction site data analysis through reasonable hierarchical design, feature extraction mechanism, activation function and regularization strategy, and can accurately process and analyze the results after multi-source information fusion, and further support key tasks such as safety assessment, equipment monitoring and progress prediction of construction sites.

[0115] like Figure 2 As shown, step 3 includes the following steps.

[0116] Conduct in-depth mining of environmental monitoring data and safety accident records from past construction projects, identify the correlation between environmental factors and the occurrence of safety accidents, and establish an environment-accident correlation model.

[0117] Compare the currently collected environmental data with the historical data model in real time, calculate the probability of safety risks occurring and generate a standardized risk score.

[0118] Based on situational awareness data, potential risks are quantitatively graded from the dimensions of personal injury, economic loss and construction delay, providing a basic basis for risk management.

[0119] Predict the risk diffusion path and impact scope, and simulate the evolution of risks in time and space dimensions.

[0120] Assess the possibility and impact of chain reactions and secondary disasters, identify key nodes and unstable links, and provide decision-making support for risk prevention and control.

[0121] Overall, the above series of steps provide scientific, efficient and comprehensive decision-making support for risk management at construction sites through deep learning, data fusion and predictive analysis, greatly improving the safety and management efficiency of construction sites.

[0122] In this embodiment, the probability of security risk occurrence is calculated by the following formula and a standardized risk assessment index is generated: score =(P(R t ∣E t )-P min ) / (P max -P min ), where Rscore It is the safety risk assessment value after standardized processing, indicating the severity and possibility of safety risks occurring at the construction site under the current environmental conditions; P min is the minimum risk probability value in historical data; P max is the maximum risk probability value in historical data; P(R t ∣E t ) is the given current environment data E t , a safety risk event R occurs t The probability of t ∣E t )=P(E t ∣R t )·P(R t ) / P(E t ), where P(E t ∣R t ) is the risk event R t Under the conditions that occur, the current environmental data E is observed t The conditional probability of t ) is the security risk event R t Prior probability of occurrence; P(E t ) is the current environment data E t The marginal probability of .

[0123] In this embodiment, predicting the diffusion path and impact range of risks and simulating the evolution of risks in time and space dimensions includes the following steps.

[0124] The following formula is used to simulate the risk propagation path at the construction site and predict the risk diffusion trend from the source point to other areas: dR g (t) / dt=∑ h∈N(g) β gh ·R h (t)+δ gh ·E h , where R g (t) is the risk value of the g-th region at time t; N(g) is the set of neighboring nodes of the g-th region, representing other locations directly related to region g; β gh is the risk transmission coefficient between the gth region and the hth region, indicating the intensity of risk transmission from region h to region g; E h is the environmental factor of the hth region; δ gh is the influence coefficient of environmental factors between the gth region and the hth region on risk transmission; R h (t) is the risk value of the hth region at time t; dR g (t) / dt is the risk change rate of the g-th region.

[0125] The following formula is used to simulate the spatial expansion of risks in different construction areas and assess the potential impact of risks: , where R(x,t) is the risk value at position x and time t; D is the risk diffusion coefficient, which indicates the diffusion speed of risk in space; ▽ 2 R(x,t) is the second-order derivative in space, which indicates the spatial expansion of risk; f(x,t) is the external source term, which indicates the impact of external factors on the spread of risk.

[0126] The following formula is used to simulate the change of risk over time and space, and analyze the impact of construction stage and environmental factors on risk evolution: R(x,t)=R0(x)·e -λt Exp(-μ‖x-x0‖), where R0(x) is the initial risk intensity, which represents the initial value at the spatial position x; -λt is the time attenuation factor, which indicates the attenuation of risk intensity with the passage of time; λ is the time attenuation coefficient, which controls the rate at which risk decreases over time; Exp(-μ‖x-x0‖) is the spatial attenuation factor, which indicates the influence of the distance from the source point x0 on the risk intensity; μ is the spatial attenuation coefficient, which controls the rate at which risk decreases with distance; ‖x-x0‖ is the distance from the spatial position x to the source point x0.

[0127] Through the following formula dR g (t) / dt=∑ h∈N(g) β gh ·R h (t)+δ gh ·E h Simulate the interactions and cascading effects between risk factors, assess how they jointly impact construction site safety, and predict potential chain reactions.

[0128] Furthermore, step 4 includes the following steps.

[0129] Establish a knowledge base of risk management solutions, organize the experience and best practices in handling various risk events in historical projects, and extract standard handling processes and key elements for different risk types.

[0130] Based on the current risk assessment results and predicted risk propagation paths, similar cases are quickly retrieved from the knowledge base, and multiple feasible disposal plans are automatically generated based on the actual situation on site.

[0131] Multi-dimensional evaluation indicators are set, including implementation cost, time efficiency, resource consumption, and safety and reliability, to quantitatively score and comprehensively rank each candidate solution.

[0132] Conduct virtual drills and effectiveness evaluations on the top-ranked disposal plans to predict the implementation effects of each plan under different conditions and potential problems.

[0133] Based on the comprehensive evaluation results, the optimal disposal plan is screened out and a detailed implementation proposal is formulated, which includes specific operating steps, required resources, personnel division of labor and emergency plans to ensure that the plan can be effectively implemented.

[0134] Overall, step 2 not only improves the accuracy and efficiency of the disposal plan through the intelligent application of historical knowledge, the support of real-time data, the introduction of multi-dimensional evaluation, and the verification of virtual drills, but also greatly enhances the construction site's ability to deal with sudden risks and its response speed.

[0135] In this embodiment, each candidate solution is quantitatively scored using the following formula: r total =w C ·(1 / C r )+w T ·T r +w R ·(1 / R r )+w S ·S r , where S r total is the comprehensive score of the rth solution; w C is the weight of implementation cost; C r Score the implementation cost of the rth option; w T is the weight of time efficiency; T r Score the time efficiency of the rth solution; w R is the weight of resource consumption; R r Score the resource consumption of the rth solution; w S is the weight of safety reliability; S r Score the safety and reliability of the rth solution.

[0136] Furthermore, step 5 includes the following steps.

[0137] Based on historical data analysis and current risk assessment results, a dynamic adjustment mechanism for warning thresholds is established, including risk level classification standards and warning trigger conditions.

[0138] Develop differentiated management strategies based on construction stages, formulate corresponding early warning standards and management requirements according to the characteristics of each stage, and automatically adjust management priorities according to the current construction progress to achieve accurate risk prevention and control.

[0139] Build a weather-sensitive early warning mechanism, access meteorological forecast data, predict the impact of weather changes on construction safety in advance, and automatically adjust the warning level and response measures according to different weather conditions to improve the targeted nature of the warning.

[0140] Continuously analyze the relationship between warning records and actual accidents, identify and eliminate false warning characteristics, gradually improve warning accuracy, and reduce resource waste caused by false alarms.

[0141] Real-time analysis and verification of abnormal data that exceeds the normal fluctuation range ensures that major risk hazards are discovered and dealt with in a timely manner, thereby achieving accurate capture and rapid response to high-risk events.

[0142] Overall, Step 5 not only improves the accuracy of warnings at the construction site through multi-dimensional risk analysis and intelligent early warning mechanisms, but also significantly enhances the ability to respond to unexpected risk events, ensuring safety and management efficiency during the construction process.

[0143] Furthermore, step 6 includes the following steps.

[0144] Based on multi-source data collection, photogrammetry and laser scanning technology are combined to perform high-precision three-dimensional reconstruction of the construction site, build a digital twin model that includes terrain, facilities and equipment elements, and update the data regularly to ensure dynamic synchronization with the site, providing a spatial reference for risk management.

[0145] The real-time collected environmental data, construction management data and risk assessment results are integrated into the digital twin model in the form of layer overlays, and the spatial distribution and change trends of various types of data are intuitively displayed through color coding or symbol annotation.

[0146] The locations of identified risk sources are marked in real time in the digital twin model, and the risk level and impact scope are displayed to achieve accurate risk management and control.

[0147] Through the playback and analysis of spatiotemporal data, we conduct in-depth research on the occurrence and development patterns of risk events, summarize experiences and lessons, and optimize risk prevention and control strategies.

[0148] The three-dimensional scene and risk monitoring functions are extended to mobile terminals, supporting managers to remotely view on-site conditions, receive risk warnings and issue emergency instructions.

[0149] Overall, Step 6 achieves comprehensive digital management of the construction site through digital twin technology and mobile support, making risk management more accurate and efficient, while improving emergency response capabilities.

[0150] like Figure 3 As shown, a highway construction safety monitoring system includes the following modules.

[0151] The data acquisition module collects environmental data, video data and construction management data from the construction site in real time, and performs preliminary processing and transmission through edge computing.

[0152] The data processing module performs spatiotemporal alignment, feature extraction and standardization on multi-source heterogeneous data, builds a unified spatiotemporal data system, and realizes deep data fusion and feature analysis.

[0153] The risk assessment module, based on historical data mining and deep learning algorithms, analyzes the correlation between environmental factors and safety accidents, evaluates risk probability and severity in real time, builds a risk propagation model, and provides support for early warning decisions.

[0154] The solution generation module retrieves and generates multiple risk disposal solutions from the knowledge base based on risk assessment results and historical experience, and optimizes sorting and screening through multi-dimensional evaluation to form an executable detailed implementation plan.

[0155] The execution monitoring module is responsible for the specific implementation of the risk disposal plan, including tracking the progress of plan execution, evaluating the execution effect and providing problem feedback, and providing real-time feedback of the execution status to relevant modules.

[0156] The early warning decision-making module dynamically formulates and adjusts early warning rules and thresholds based on seasonal, weather and construction stage factors and combined with risk assessment results to achieve hierarchical early warning and differentiated management strategies.

[0157] The intelligent interactive module visualizes construction site data and risk assessment results through 3D reconstruction and digital twin technology, and supports risk source location, historical backtracking, and remote monitoring and command on mobile terminals.

[0158] The highway construction safety monitoring system integrates a number of advanced technologies to effectively improve the safety monitoring and risk management capabilities of the construction site. First, the data acquisition module collects environmental data, video data and construction management data in real time, and uses edge computing for preliminary processing and transmission to ensure the real-time and stability of the data. The data processing module performs spatiotemporal alignment, feature extraction and standardization of multi-source heterogeneous data, establishes a unified spatiotemporal data system, and ensures the deep integration and comprehensive analysis of data. The risk assessment module analyzes the correlation between environmental factors and safety accidents through historical data mining and deep learning algorithms, evaluates the risk probability and severity in real time, and constructs a risk propagation model to provide accurate support for early warning and decision-making. The solution generation module uses historical experience in the knowledge base and combines risk assessment results to automatically generate multiple risk disposal solutions, and optimizes and sorts them through multi-dimensional evaluation to ensure that the proposed disposal solutions are both feasible and efficient. The execution monitoring module is responsible for implementing the risk disposal plan, tracking the execution progress, evaluating the execution effect and providing problem feedback to ensure the effective implementation of the plan. The early warning decision module dynamically adjusts the early warning rules and thresholds according to the season, weather and construction stage factors of the construction site, implements graded early warning and differentiated management strategies, and improves the accuracy and pertinence of early warning. The intelligent interactive module visualizes the construction site data and risk assessment results through three-dimensional reconstruction and digital twin technology, provides risk source location, historical backtracking, and mobile remote monitoring and command functions, greatly enhancing the real-time and flexibility of on-site management and decision-making.

[0159] In summary, the system provides efficient and accurate monitoring and decision-making support for highway construction safety management through the combination of data fusion, deep learning, risk assessment, dynamic adjustment and visualization technology, ensuring safety and efficiency during the construction process.

[0160] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A highway construction safety monitoring system, characterized in that: Includes the following modules: Data acquisition module, which collects environmental data, video data and construction management data of the construction site in real time; Data processing module, building a unified spatiotemporal data system to achieve deep data integration and feature analysis; The risk assessment module analyzes the correlation between environmental factors and safety accidents, and assesses the risk probability and severity in real time; The solution generation module generates multiple risk treatment solutions and optimizes and sorts them through multi-dimensional evaluation to form an executable detailed implementation plan; The execution monitoring module is responsible for the specific implementation of the risk disposal plan, including plan execution progress tracking, execution effect evaluation and problem feedback, and provides real-time feedback of the execution status to relevant modules; Early warning decision module, realizing graded early warning and differentiated management strategies; The intelligent interactive module visualizes construction site data and risk assessment results, and supports risk source location, historical backtracking, and mobile remote monitoring and command.

2. A highway construction safety monitoring method, implemented based on the highway construction safety monitoring system according to claim 1, characterized in that: The following steps are involved: Step 1: Real-time collection of construction site data, including environmental data, video data, and construction management data; Step 2, generating global situation awareness data; Step 3: Assess the probability and severity of safety risks in real time, and predict the risk propagation path and cascading effects; Step 4: Generate multiple risk treatment plans, sort and screen them, and select the best treatment suggestion; Step 5: Achieve accurate early warning and differentiated management, reduce false alarm rates and detect high-risk events in a timely manner; Step 6: Display the construction site through 3D reconstruction, overlay the construction site data and risk assessment results, and provide mobile terminal remote monitoring and emergency command.

3. A highway construction safety monitoring method according to claim 2, characterized in that: Step 1 includes the following steps: Deploy multiple environmental monitoring units at the construction site; Install high-definition cameras and intelligent monitoring equipment at key locations on the construction site to record the construction process in real time; Obtain construction progress, staffing, equipment usage and material consumption data from the construction management system; Preliminary screening, cleaning and classification of the collected raw data to eliminate abnormal data; Establish a data backup mechanism and regularly back up all collected data.

4. A highway construction safety monitoring method according to claim 2, characterized in that: Step 2 includes the following steps: Perform time stamp calibration and synchronization on the collected environmental monitoring data, video data and construction management data; By accurately locating monitoring points, camera locations, and construction areas, spatial correlation of data is achieved, and a complete spatiotemporal data system is constructed; Extract features from video data to identify key objects, events, and behavior patterns at the construction site; Conduct multi-dimensional feature analysis on environmental monitoring data, extract temperature change trends, noise fluctuation patterns and air quality index characteristics, establish correlation models between environmental factors, and quantitatively describe environmental trends; The extracted features are deeply integrated to build a multi-source information fusion model and generate standardized situation awareness data.

5. A highway construction safety monitoring method according to claim 4, characterized in that: The extracted features are deeply integrated to build a multi-source information fusion model to generate standardized situation awareness data, including the following steps: Perform target detection and tracking on construction site videos, identify the location and dynamic information of construction personnel, equipment, and materials, perform semantic segmentation on construction scenes, and store the extracted feature data in a unified standard format; Clean and standardize the environmental feature data, construct time series features, and generate an environmental feature map of the construction site in combination with the sensor location information; Structural processing of construction management data, establishment of unified data standards, and conversion of unstructured text information into a quantifiable indicator system; Establish a unified indexing system based on timestamps and spatial coordinates to align video features, environmental features, and management data in time and space dimensions; The aligned multi-source feature data is input into the deep neural network model, and a unified feature representation is generated through multi-level feature extraction and fusion, and the corresponding output layer is designed according to the specific needs of the construction site; The output results of the deep neural network model are post-processed to form a comprehensive situational awareness dataset that includes personnel safety assessment, equipment operating status, environmental condition evaluation, and construction progress quantitative indicators.

6. A highway construction safety monitoring method according to claim 2, characterized in that: Step 3 includes the following steps: Conduct in-depth mining of environmental monitoring data and safety accident records from past construction projects, identify the correlation between environmental factors and safety accidents, and establish an environment-accident correlation model; Compare the currently collected environmental data with the historical data model in real time, calculate the probability of safety risks and generate standardized risk scores; Based on situational awareness data, potential risks are quantitatively graded from the dimensions of personal injury, economic loss and construction delay; Predict the diffusion path and impact range of risks, and simulate the evolution of risks in time and space dimensions; Assess the possibility and impact of chain reactions and secondary disasters, and identify key nodes and unstable links.

7. A highway construction safety monitoring method according to claim 2, characterized in that: Step 4 includes the following steps: Establish a risk management solution knowledge base, organize the experience and best practices in handling various risk events in historical projects, and extract standard handling processes and key elements for different risk types; Based on the current risk assessment results and predicted risk propagation paths, similar cases can be quickly retrieved from the knowledge base, and multiple feasible disposal plans can be automatically generated based on the actual situation on site; Quantitatively score and comprehensively rank each candidate solution, conduct virtual drills and effect evaluations on the top-ranked disposal solutions, and predict the execution effects of each solution under different conditions and potential problems; Based on the comprehensive evaluation results, the optimal disposal plan is screened out and a detailed implementation proposal is formed.

8. A highway construction safety monitoring method according to claim 2, characterized in that: Step 5 includes the following steps: Establish a dynamic adjustment mechanism for early warning thresholds based on historical data analysis and current risk assessment results; Formulate corresponding early warning standards and management requirements according to the characteristics of each stage, and automatically adjust management priorities according to the current construction progress to achieve accurate risk prevention and control; Build a weather-sensitive early warning mechanism to automatically adjust the warning level and response measures according to different weather conditions; Continuously analyze the correlation between warning records and actual accidents, identify and eliminate false warning features; Conduct real-time analysis and verification of abnormal data that exceeds the normal fluctuation range to ensure that major risks and hidden dangers are discovered and dealt with in a timely manner.

9. A highway construction safety monitoring method according to claim 2, characterized in that: Step 6 includes the following steps: Carry out high-precision 3D reconstruction of the construction site, build a digital twin model that includes terrain, facilities, and equipment elements, and regularly update the data to ensure synchronization with the site dynamics; Integrate the real-time collected environmental data, construction management data and risk assessment results into the digital twin model in the form of layer overlay, and display the spatial distribution and change trend of various data; Mark the location of identified risk sources in the digital twin model in real time, and display the risk level and impact range; Conduct in-depth research on the occurrence and development patterns of risk events, summarize lessons learned, and optimize risk prevention and control strategies; The three-dimensional scene and risk monitoring functions are extended to mobile terminals, supporting remote viewing of on-site conditions, receiving risk warnings and issuing emergency instructions.

Images