CN117788223A

CN117788223A - Building construction management method based on multi-data fusion

Info

Publication number: CN117788223A
Application number: CN202410085647.2A
Authority: CN
Inventors: 马行耀; 叶佳赟; 齐琳
Original assignee: Zhejiang College of Construction
Current assignee: Zhejiang College of Construction
Priority date: 2024-01-22
Filing date: 2024-01-22
Publication date: 2024-03-29

Abstract

The invention relates to a building construction management method based on multi-data fusion. Firstly, a real-time multi-source data integration and analysis platform is adopted, which is used for integrating multi-source data of a building construction site, including material tracking, worker positions and machine states, into a unified platform, and processing information of various data sources from Internet of things equipment, unmanned aerial vehicle monitoring and personnel positioning systems through a data processing engine; secondly, adopting an intelligent prediction and risk assessment model, analyzing historical data and real-time data by utilizing machine learning and artificial intelligent technology to predict risk and construction progress, wherein the method comprises the steps of collecting the historical construction data to train the prediction model, and inputting new data in real time to perform risk assessment and progress prediction; in addition, the method also comprises a self-adaptive resource optimization scheduling algorithm, and the allocation of resources such as manpower, machinery and materials is automatically adjusted according to real-time data so as to optimize the construction efficiency, and the allocation of the resources is dynamically adjusted according to real-time scenes and predicted data.

Description

Building construction management method based on multi-data fusion

Technical Field

The invention relates to a building construction management method, in particular to a building construction management method based on multi-data fusion.

Background

The shortcomings and drawbacks of the current methods for building construction management mainly include information isolation and data fragmentation, lack of real-time data and analysis capabilities, insufficient risk management, security issues, environmental protection issues, cost and budget control issues, communication and coordination difficulties, slow adoption of technologies and innovations, insufficient labor management, insufficient response to change management, quality control issues, and insufficient document and record management.

In the aspects of information isolation and data fragmentation, a plurality of departments and teams use different systems and tools to record and communicate data, so that the information transmission efficiency is low, and the effective sharing and comprehensive analysis of the data cannot be realized. This fragmented data management approach results in a slow and error-prone management decision process that affects the progress and quality of the overall project. Lack of real-time data and analysis capability is another important issue. In many traditional construction projects, project progress and monitoring of resource usage relies on periodic reporting and manual inspection to provide no real-time data, resulting in a management team failing to quickly respond to site changes, possibly resulting in excessive or insufficient use of resources.

Inadequate risk management is also a significant problem. In the conventional method, risk assessment mainly depends on experience judgment and qualitative analysis, and lack of a system risk identification and quantitative management mechanism may lead to neglect or misjudgment of potential risks. Safety issues are particularly important in building construction. In traditional construction management, safety monitoring relies on manual inspection and self-management of workers, and an effective technical means is lacked for comprehensive monitoring and early warning. Environmental protection issues are not neglected. Many conventional building construction management methods fail to adequately account for environmental impact, which may lead to violation of environmental regulations.

Cost and budget control are key aspects of building construction management. However, conventional approaches often have difficulty accurately predicting and controlling costs. Communication and coordination obstacles are challenges presented by the collaboration of multiple professional teams. The traditional management method has barriers in communication and coordination, especially when information updating is not timely or information transmission is error. The adoption of technology and innovation has slowly limited the efficiency and effectiveness of construction management. Insufficient labor management is also a problem, and traditional construction management fails to effectively manage labor resources. Insufficient response to change management is also a problem, changes are often encountered in the building construction process, and conventional management methods often have insufficient timely and flexible response to such changes.

Finally, the quality control and document management deficiency are also drawbacks of the traditional building construction management method. In conventional construction management, quality control often relies on post-inspection rather than overall process quality assurance. The lack of an efficient document management system also makes the traceability and sharing of project information difficult.

Disclosure of Invention

The invention aims to provide a building construction management method based on multi-data fusion, so as to solve part of defects and shortcomings pointed out in the background art.

The invention solves the technical problems as follows: the method comprises the following steps:

firstly, a real-time multi-source data integration and analysis platform is adopted, which is used for integrating multi-source data of a building construction site, including material tracking, worker positions and machine states, into a unified platform, and processing information of various data sources from Internet of things equipment, unmanned aerial vehicle monitoring and personnel positioning systems through a data processing engine;

secondly, adopting an intelligent prediction and risk assessment model, analyzing historical data and real-time data by utilizing machine learning and artificial intelligent technology to predict risk and construction progress, wherein the method comprises the steps of collecting the historical construction data to train the prediction model, and inputting new data in real time to perform risk assessment and progress prediction;

The method also comprises a self-adaptive resource optimization scheduling algorithm, wherein the allocation of resources such as manpower, machinery and materials is automatically adjusted according to real-time data so as to optimize the construction efficiency, and the allocation of the resources is dynamically adjusted according to real-time scenes and predicted data;

the intelligent safety monitoring system is combined, the safety condition of a construction site is monitored in real time by adopting a machine vision and sensing technology, and potential hazards are timely alerted, and the system comprises a high-resolution camera and a sensor which are deployed, and a software system for detecting potential safety threats is developed;

and finally, adopting an environmental impact assessment tool for assessing and minimizing the impact of construction on the environment, including noise, dust and waste management, analyzing the impact of construction activities on the surrounding environment by creating a model, and formulating a slowing measure.

Further, the real-time multi-source data integration and analysis platform comprises:

s1, firstly, constructing a modularized and expandable data integration architecture, and applying a data integration function:

the method is directly applied to data fusion of different data sources including a material tracking system, worker position tracking and machine state monitoring, wherein x, y and z respectively represent key parameters of the different data sources, and N is a parameter for adjusting data depth fusion;

S2, then adopting real-time data stream processing, and passing through the formula:

processing real-time data streams from the monitoring of the Internet of things equipment and the unmanned aerial vehicle, wherein lambda represents a data stream dynamic adjustment factor; and then, real-time acquisition of physical data on site by utilizing the Internet of things equipment, and a data acquisition formula:

optimizing a data acquisition process, wherein d represents a multidimensional distance between data points, and p and q represent acquisition precision and quality;

s3, unmanned aerial vehicle visual angle data are applied, site conditions are analyzed through image recognition and machine learning algorithms, and an image analysis function is applied:

H(i，j，k)＝ln(i ^αk +e ^jβ )

wherein i, j and k respectively represent pixel intensity, position parameters and time sequence of the image, and alpha and beta are adjustment parameters;

s4, finally, displaying analysis results through a data analysis and three-dimensional data visualization interface, and using a function:

and adjusting the visual display of the data, wherein r represents the data correlation, s and t represent the multidimensional scale of the data display.

Further, the intelligent prediction and risk assessment model comprises the following steps:

firstly, data preprocessing and integration are carried out, and a data fusion function is applied:

to process data x, y, z from different sources including construction logs, weather records, resource usage;

And then executing feature engineering, and extracting a function through the features:

extracting key indexes from historical data, wherein a and b represent characteristic values in an original data set;

then developing a machine learning model suitable for construction management, and adopting a model optimization formula:

wherein p is _i And q _i Respectively representing model parameters and corresponding weights;

and finally, fusing the real-time data into a model, and using a real-time data updating function:

I(r，s)＝∫e ^r·s dr

where r represents the real-time data point and s represents the update frequency.

Further, the adaptive resource optimization scheduling algorithm comprises the following steps: firstly, multidimensional data integration is carried out, and a data fusion function is applied:

processing real-time information of manpower distribution, machine use state and material stock level from a construction site, wherein x, y and z respectively represent different data dimensions;

and then, performing real-time monitoring and analysis, and performing a real-time data analysis formula:

G(a，b，t)＝a·e ^-bt

continuously collecting and processing field data, wherein a, b is an adjustment parameter, and t is a time variable;

then developing a prediction model driven by machine learning or artificial intelligence, and adopting a prediction model optimization formula:

wherein p represents model parameters, q represents input data to realize optimization and scheduling of construction resources; the gradient of the parameter p in the logistic regression model is calculated, so that the parameter is updated according to gradient information in the training process, and continuous optimization and accurate prediction of the model are realized; and meanwhile, the specific form of q is adjusted according to the data set and the prediction task, the method is suitable for application scenes, and the generalization capability and the prediction accuracy of the model are improved by combining with other machine learning technologies including feature engineering and regularization strategies.

Further, the safety monitoring method realized by combining the machine vision and the sensing technology comprises the following steps: the method comprises the steps of utilizing a machine vision algorithm developed by deep learning and image processing technology to analyze image and video data captured by a high-resolution camera in real time so as to identify potential safety hazards including illegal invasion and unworn safety helmet;

second analyzing data from various sensors including temperature, humidity, vibration to detect risks including fire or structural instability;

the camera and sensor data are comprehensively analyzed, and various potential risks are predicted and identified by utilizing an artificial intelligence algorithm, so that real-time alarm when danger is detected is realized;

finally, the system comprises a comprehensive software system, integrates visual and sensing data, and realizes real-time processing, risk assessment and alarm notification of the data.

Further, the real-time analysis of the image and video data captured by the high-resolution camera uses:

s1, combining deep learning and image processing technology, and applying an image processing function:

preprocessing the data captured by the camera, enhancing the image quality, and improving the recognition accuracy of a subsequent deep learning model;

S2, adopting a deep learning model to utilize an optimization function:

to accurately identify a particular security risk, where x _n A parameter set representing the model, y representing the input image data, and N representing the number of parameters;

s3, adopting an integrated real-time analysis algorithm:

the camera and sensor data are analyzed, potential risks are timely identified and alarmed, wherein t represents a time variable, and s represents a data sensitivity factor.

Further, the machine vision algorithm developed by adopting the deep learning and image processing technology is used for comprehensively analyzing various sensor data including temperature, humidity and vibration; by applying the anomaly detection formula:

wherein x is _i Represents sensor data points, μ _i Sum sigma _i Respectively representing the mean value and standard deviation of the data to identify abnormal modes in the data;

and finally, adopting a risk identification model optimization formula:

where p represents the sensor dataset and q represents the risk threshold for predicting the risk of e.g. fire or structural instability.

Further, the method for predicting and identifying various potential risks by the artificial intelligence algorithm comprises the following steps:

firstly, analyzing camera and sensor data by using machine learning and deep learning algorithms, and adopting a feature extraction formula:

Wherein x is _i And y _i Representing features extracted from the image and sensor data, respectively;

secondly, implementing a real-time monitoring and early warning system, and using a real-time data analysis function:

G(t)＝∫ ₀ ^t s·e ^-λs ds

where t represents time, s represents sensor signal strength, lambda is a decay factor to analyze the data stream and alert when an abnormal or dangerous pattern is detected.

Further, the environmental impact assessment tool comprises:

s1, firstly, using an environmental impact model:

wherein x, y, z represent the measurements of noise, dust and waste, respectively, a, b, c are impact weights to analyze the impact of construction activities on the surrounding environment; the integral is the sum of the environmental impact of these factors over a range from a to b;

s2, applying a machine learning algorithm, and adopting a risk identification formula:

wherein d is _i Representing features extracted from the environmental data for predicting environmental risk; the risk identification formula calculates the contribution of the features to the overall environmental risk and is used for converting the data features into risk assessment;

s3, then, making a slowing measure according to the model analysis result, and utilizing a slowing effect evaluation function:

wherein t represents time, u represents the effect of the slow-down measure, and alpha is an adjustment coefficient; the slowing effect evaluation function combines the influence of the change rate of the slowing effect along with time and the adjusting coefficient on the effect;

S4, finally, continuously monitoring and adjusting a mechanism, and applying an adjusting feedback function:

where v represents an environmental impact variable and λ is an adjustment factor to ensure environmental impact minimization and practicality; the adjustment feedback function formula represents the rate of change of the environmental impact variable over time for guiding the adjustment of the environmental impact assessment.

The invention has the following beneficial effects:

1. the decision quality and the efficiency are improved: by comprehensively analyzing information from cameras, sensors, and other data sources, this approach provides a comprehensive insight to the management team, enabling them to make more accurate and rapid decisions based on data-driven insights.

2. Enhancing risk management: by utilizing the deep learning and machine learning algorithms, the team can predict and identify various potential risks, such as safety accidents, fire disasters or unstable structures, and the like, so that preventive measures can be taken in time, and the probability of accident occurrence is reduced.

3. Optimizing resource allocation: by monitoring and analyzing the use conditions of manpower, machinery and materials in real time, the method can optimize the allocation and use of resources, improve the construction efficiency and reduce the cost.

4. And (3) improving the safety standard: the real-time analysis of the video and the sensor data is helpful for timely finding potential safety hazards, such as unworn safety helmets, illegal invasion and the like, so that the safety of on-site staff is enhanced.

5. Environmental protection: by evaluating the impact of construction activities on the environment and taking corresponding mitigation measures, this approach helps to reduce negative impact on the surrounding environment, promoting sustainable development.

Drawings

FIG. 1 is a full flow chart of a construction management method based on multi-data fusion.

FIG. 2 is a flow chart of a security monitoring method implemented in combination with machine vision and sensing techniques in accordance with the present invention.

Detailed Description

The following describes the embodiments of the present invention in detail with reference to the drawings.

The method comprises the step of adopting a real-time multi-source data integration and analysis platform, wherein the platform aims at integrating multi-source data of a building construction site, such as material tracking, worker position, machine state and the like, and integrating the data of different sources into a unified platform. Such integration is achieved through an efficient data processing engine capable of processing information from a variety of data sources, including internet of things devices, unmanned aerial vehicle monitoring systems, personnel location systems, and the like. In this way, the platform can provide a comprehensive view reflecting the real-time conditions of the construction site. This not only helps to improve resource management and efficiency, but also may enhance field safety monitoring and workflow optimization. Integrating these data enables the manager to make faster, more accurate decisions.

The second key step of the scheme is to adopt an intelligent prediction and risk assessment model. This step involves analyzing historical and real-time data using machine learning and artificial intelligence techniques to predict potential risks and construction progress. This process first involves collecting historical construction data that is used to train a machine learning model to enable it to identify and understand the various patterns and trends in the construction process. Once the model is trained, it can be used to analyze the new data entered in real time for more accurate risk assessment and progress prediction. This approach enables construction managers to make decisions based on data-driven insights, identify and address issues in advance, thereby optimizing construction plans, improving efficiency, and reducing risk.

The scheme also comprises an adaptive resource optimization scheduling algorithm, which is a key function, so that the construction management system can automatically adjust resource allocation according to real-time data. These resources include manpower, machinery, materials, etc., the allocation and utilization of which is critical to the efficiency of construction. The algorithm dynamically optimizes resource allocation by analyzing data and predictive information of the real-time scene. For example, if the forecast data indicates that a particular work area will require more human resources, the algorithm may automatically adjust the allocation of workers to meet the upcoming demand. Also, if the real-time data indicates that a certain material is being used faster than expected, the algorithm may adjust the supply of material in time to avoid operational delays.

The method further comprises the step of combining an intelligent safety monitoring system which adopts advanced machine vision and sensing technology to monitor the safety condition of a construction site in real time and give an alarm in time when a potential danger is found. The system includes the deployment of high resolution cameras and a variety of sensors that are capable of capturing a variety of information on site, including personnel activities, machine operating conditions, and other safety hazards. In addition, specialized software systems have been developed to analyze such data, which utilize algorithms to detect potential security threats, such as illegal intrusions, security protocol violations, and the like. The intelligent safety monitoring system greatly improves the safety management level of a construction site, effectively prevents accidents through a real-time monitoring and quick response mechanism, and ensures the safety of workers and equipment.

The final step of the scheme comprises the adoption of an environmental impact assessment tool, and aims to assess and minimize the impact of construction activities on the environment, particularly concerning noise, dust, waste management and the like. This tool analyzes how construction activities affect the surrounding environment, including increases in noise levels, changes in air quality, and waste generation and disposal, by creating models. Based on the analysis results of this model, the management team can formulate specific mitigation measures, such as employing noise reduction techniques, improving dust control strategies, and optimizing waste treatment processes. The use of the environmental impact assessment tool enables the construction project to not only effectively manage the impact on the environment, but also meet environmental protection standards and regulations, thereby being beneficial to realizing sustainable construction targets and protecting natural environment and living communities.

Example 1: building construction projects of a certain large office building. Project management teams decide to adopt a building construction management method based on multi-data fusion to improve efficiency and safety. As a core part of this approach, the present embodiment begins to build a real-time multi-source data integration and analysis platform.

Step S1, constructing a modularized and extensible data integration architecture:

first, project teams develop a modular, scalable data integration architecture. This architecture is designed to process and fuse various data from the job site. For example, the material tracking system provides data regarding the use and location of construction materials, the worker position tracking system provides the exact location of each worker in the field, and the machine condition monitoring system feeds back the operating conditions of the various construction machines in real time.

To integrate these different data sources, teams employ specific data integration functions:

in this function, x, y, z represent key parameters collected from different data sources, respectively, such as x representing the quantity of material, y representing the worker's position coordinates, and z representing the machine's length of operation. N is an adjustment parameter for controlling the depth and complexity of the data fusion. In certain construction scenarios, the team needs to integrate the following data: the material tracking system showed 50 cubic meters of concrete (x=50), 5 workers were in a specific area (y=5), and a certain excavator had been operated for 8 hours (z=8). And a specific value can be obtained by applying the data integration function and is used for evaluating the resource allocation and the use efficiency in the current scene. Through such data integration and analysis, the project management team can more effectively monitor the progress of construction, ensure optimal utilization of resources, and adjust the plan in time to cope with various situations. The method ensures that the construction management is more intelligent and automatic, and greatly improves the efficiency and the safety.

Step S2, real-time data stream processing:

to more efficiently process and analyze real-time data streams from internet of things devices and drone monitoring, teams employ specialized data stream processing formulas:

in this formula λ represents a data stream dynamic adjustment factor for balancing and adjusting the processing rate and quality of the data stream. t represents time and v represents the rate or number of data streams. The purpose of this formula is to optimize the processing of real-time data, ensuring that the data analysis is both timely and accurate. For example, at a specific time point, the sensors of the internet of things report significant changes in data such as temperature, humidity, vibration, etc. of the construction site, and the unmanned aerial vehicle monitors and captures some abnormal activities. The data is fed into a data stream processing formula where lambda is set to a specific value to ensure that the data reflects critical security information while being efficiently processed. Meanwhile, the team also collects physical data on site in real time by using the Internet of things equipment. To optimize this process, the present embodiment applies the data acquisition formula:

here, d represents the multidimensional distance between the data points, and p and q represent the acquisition accuracy and quality, respectively. By this formula, the team can more precisely determine which data points are most important and adjust the acquisition strategy of the present embodiment to obtain the most critical information. The introduction of the data stream processing and optimizing collection strategies obviously improves the real-time knowing and responding capability of the team to the condition of the construction site. The processing of emergency events and the efficiency of daily management are improved remarkably.

Step S3, unmanned aerial vehicle visual angle data are applied:

the team decides to utilize the high definition images captured by the drone to further monitor and analyze the job site. To efficiently analyze these image data, the present embodiment employs a specific image analysis function:

H(i，j，k)＝ln(i ^αk +e ^jβ )

in this formula, i, j, k represent the pixel intensity, the position parameter, and the time series extracted from the drone image, respectively. Alpha and beta are tuning parameters for adjusting the sensitivity and accuracy of the image analysis. This function enables a team to more accurately identify important features in the image, such as construction progress, personnel distribution, and potential safety hazards.

As a specific example, the drone captures an image of a set of construction areas, a portion of which shows a new piece of concrete area laid. By applying the image analysis function, the team can calculate the eigenvalues of these areas. In a particular image, the pixel intensity i is 200, the position parameter j is 50, the time series k is 3 hours, and the adjustment parameters α and β are set to 0.5 and 2, respectively. Substituting these values into a formula, the team can get a specific value that helps the present embodiment understand the specific condition of the area, such as whether concrete is properly laid, or whether any anomalies are occurring. The application of the unmanned aerial vehicle visual angle data provides a brand new visual angle for project management teams to monitor and evaluate construction sites by combining with advanced image analysis technology. The method not only improves the monitoring range and efficiency, but also enhances the control capability of the team on the construction progress and site safety.

Step S4, data analysis and three-dimensional data visualization interface display:

team decision uses a special function to optimize visual presentation of data:

in this formula, r represents the correlation of the data, and s and t represent the multidimensional scale of the data presentation. The purpose of this function is to present the data in an intuitive and informative way, enabling a team to quickly understand and respond to various conditions at the job site. For example, a team may wish to demonstrate worker distribution and the operating status of a construction machine. The correlation r of this embodiment is 0.8 from the data stream processing, indicating a strong correlation between worker distribution and machine operation. To present these data in three-dimensional space, the present embodiment sets s and t to represent the extent of worker distribution and the frequency of machine operation, s=5 and t=3, respectively. Substituting these values into the formula results in a specific value that is used to adjust the dimensions and viewing angle of the three-dimensional visualization chart, making the data presentation both intuitive and informative.

Through such three-dimensional data visualization, project management teams can see visual representations of complex data, such as worker profiles, machine use, and other critical construction indicators, on visual interfaces. This visualization approach not only makes the data easier to understand, but also facilitates faster decision-making and more efficient resource allocation. In general, by the construction management method based on the multi-data fusion, a team can more effectively process and analyze a large amount of construction data, thereby realizing more efficient and safer construction management.

Example 2: in a large office building construction project, in order to more accurately predict risks occurring in a construction process and evaluate the influence of the risks on the overall progress, a project management team decides to adopt an intelligent prediction and risk evaluation model. The first step of this model is data preprocessing and integration.

Step 1, data preprocessing and integration:

teams first integrate data from different sources, including construction logs, weather records, resource usage, and the like. To efficiently process these different types of data, the present embodiment employs a data fusion function:

in this formula, x, y, z represent key parameters from different data sources, respectively. For example, x represents a temperature record of a day, y represents the amount of rainfall on the same day, and z represents the amount of usage of a particular resource.

In a specific scenario, the construction log shows that there is rainfall (y=10 millimeters) at high temperatures of one day (x=35 degrees celsius), and a lot of concrete is consumed (z=100 cubic meters). By substituting these values into the data fusion function, the management team can calculate a specific value that helps this embodiment understand the efficiency of resource usage under specific weather conditions. In this way, a team can convert data that is otherwise scattered and difficult to compare into a more uniform and analyzable format. This not only allows the present embodiment to better understand the relationships between data, but also provides a solid basis for subsequent risk assessment and prediction.

Step 2, executing characteristic engineering:

the core of this step is to extract key indicators from the historical data, and to achieve this goal, the team uses feature extraction functions:

in this formula, a and b represent eigenvalues in the original dataset. For example, a represents the average number of hours of operation per day over the past week, and b represents the average material consumption per day over the same period of time. Within a certain week, the team recorded an average of 10 hours of work per day (a=10) and an average of 50 cubic meters of concrete consumed per day (b=50). Substituting these values into the feature extraction function, specific indices reflecting the relationship between the operating time and the material consumption and the trend of the relationship over time can be calculated. In this way, a team can understand and quantify the dynamic relationship between work efficiency and resource usage in depth.

These insights are critical to predicting future resource demands, optimizing work schedules, and identifying resource shortages or waste in advance. Such feature engineering not only makes the historical data more analytical, but also provides a solid basis for making more accurate predictions using machine learning models.

Step 3, developing a machine learning model:

the team aims to create a model, so that the construction progress and potential risks can be accurately predicted, and more efficient and safe construction management is realized. To optimize this machine learning model, the present embodiment employs a model optimization formula:

in this formula, p _i And q _i Representing the parameters of the model and the corresponding weights, respectively. These parameters cover various metrics extracted from historical data and real-time data, such as worker's work efficiency, material consumption rate, machine usage, etc.

During the model development process, teams recognize several key parameters, such as daily concrete usage (p ₁ ) Average working time of worker (p ₂ ) And failure rate of machine (p ₃ ). Accordingly, the present embodiment assigns different weights (q ₁ ，q ₂ ，q ₃ ) To reflect their extent of impact on the progress of the construction. By substituting these parameters and weights into the optimization formula, the team can adjust and optimize the behavior of the model, ensuring that it can accurately predict key variables and potential risks in the construction process. The development of the machine learning model not only improves the prediction accuracy, but also enhances the team control on the construction project. The project management team can recognize problems in advance and make adjustments quickly, so that the waste of cost and time is avoided, and the smooth construction is ensured.

Step 4, fusing the real-time data into a model:

to achieve this goal, the team employs a real-time data update function:

I(r，s)＝∫e ^r·s dr

in this formula, r represents a real-time data point collected from a construction site, such as a material consumption amount or a real-time working state of a worker at a specific time point, and s represents a frequency of data update. For example, at a construction site, a team monitors the amount of concrete used and the activity of workers in real time through sensors. At a specific moment, the amount of concrete used is r=60 cubic meters, and the data update frequency s is set to once per hour. The team substitutes these real-time data points into the update function to calculate a specific value that helps the present embodiment update the relevant parameters in the machine learning model, ensuring that the model reflects the latest job site situation. In this way, the machine learning model can adapt to changes in the construction site in real time, more accurately predicting and assessing risk. The method not only improves the efficiency and accuracy of risk management, but also provides real-time feedback for project management teams, so that the method can adjust construction strategies and resource allocation in time.

In general, through the intelligent prediction and risk assessment model based on multi-data fusion, a construction management team can control project progress more effectively, accidents and risks are reduced, resource utilization is optimized, and accordingly smooth performance of building projects according to plans is ensured.

Example 3: in the construction project of a large office building, a project management team adopts a self-adaptive resource optimization scheduling algorithm to effectively manage manpower, machinery and material resources of a construction site. The first step of this algorithm is to perform multidimensional data integration.

First step, multidimensional data integration:

the team's goal is to monitor and optimize the resource allocation at the job site in real time. For this purpose, the present embodiment first integrates various data from the construction site, including real-time information such as human distribution, machine use status, and stock level. To efficiently process these different dimensions of data, the present embodiment employs a data fusion function:

in this formula, x, y, z represent the values of the different data dimensions, respectively. For example, x may be the number of workers in a particular area, y may be the run time of the machine in that area, and z may represent the amount of material used.

In a specific scenario, the team measures that 15 workers are working in a work area (x=15), the machinery in that area has been in operation for 6 hours (y=6), and 30 cubic meters of concrete are consumed (z=30). By substituting these data into the fusion function, the team can calculate a specific value that helps the present embodiment understand the current resource allocation situation and evaluate whether adjustments are needed.

By the multidimensional data integration method, teams can more comprehensively know the actual conditions of the construction site. This not only helps this embodiment to more effectively utilize resources, but also enables this embodiment to timely adjust the distribution of manpower and machinery, optimizing the use of materials, thereby improving overall construction efficiency.

Secondly, monitoring and analyzing in real time:

to effectively perform real-time monitoring and analysis, the team employs a real-time data analysis formula:

G(a，b，t)＝a·e ^-bt

in this formula, a and b are tuning parameters for adjusting the sensitivity and response rate of the data analysis, and t is a time variable representing the length of time from the beginning of data collection to the present. For example, a team is monitoring a concrete placement operation in a critical area. To evaluate the rate of progress of the work, the present example sets a to the initial work rate, such as cubic meters of poured concrete per hour, to a = 20 cubic meters per hour. b is set to a coefficient reflecting the change in the operation efficiency, b=0.05. Over time (e.g., t is 5 hours), the team can calculate the current work efficiency using this formula. By substituting these values into the formula, work progress data updated in real time can be obtained, which is crucial for timely adjustment of resource allocation. If the results of this formula show a decrease in the work rate (i.e., a slower concrete placement rate), the team needs to add manpower or adjust the work method.

Thirdly, developing a prediction model:

team starts to develop machine learning or artificial intelligence driven predictive models aimed at further optimizing allocation and scheduling of construction resources. To achieve this objective, the present embodiment employs a predictive model optimization formula:

in this formula, p represents a model parameter, and q represents input data. This formula is essentially a gradient expression of a logistic regression model for calculating the gradients of the model parameters and updating the parameters according to these gradient information during the training process. For example, a team is predicting the concrete demand in the next week. In this embodiment, the model parameter p is set to historical concrete usage data, and q is a specific feature of the current construction plan. By substituting these data into the formula, the present embodiment can calculate the gradient of the model parameters and adjust the prediction model accordingly to more accurately reflect the future demands. In addition, the team adjusts the specific form of q according to the specific data set and the prediction task so as to adapt to different application scenes. The present embodiment also combines this predictive model with other machine learning techniques, including feature engineering and regularization strategies, to improve the generalization ability and predictive accuracy of the model. By the method, the team can more accurately predict the demand of construction resources, so that the optimization and effective scheduling of the resources are realized. The construction efficiency is improved, the resource waste is reduced, and the construction progress is ensured to be consistent with the plan.

Example 4: in a large office building construction project, in order to improve the safety management level of a construction site, a project management team decides to develop a machine vision algorithm by using deep learning and image processing technology. The goal of this algorithm is to analyze the image and video data captured by the high resolution camera in real time to identify potential security hazards including illegal intrusion, non-wearing of the helmet, etc.

S1, image preprocessing:

first, a team pre-processes the data captured by the camera using an image processing function. This step is critical because high quality image data is the basis for ensuring accuracy of recognition of subsequent deep learning models. The image processing function is defined as:

the function is to denoise the captured image and enhance contrast, thereby improving the quality of the image.

For example, the drone captures a set of images of the construction area, some of which appear blurred due to poor lighting conditions. The team inputs these image data into the image processing function. In the processing process, the function reduces noise by adjusting the contrast and brightness of the image, so that the image is clearer.

Through such preprocessing, important features in the image, such as the position of personnel, whether to wear a safety helmet, etc., become more easily identified by a subsequent deep learning model. This not only improves the accuracy of the model, but also provides more reliable data for teams to monitor and risk assessment. By applying the method combining the deep learning and image processing technology, project management teams can more effectively identify and prevent potential safety hazards of construction sites, thereby ensuring the safety of staff and the smooth progress of construction progress.

S2, deep learning model application:

project teams employ deep learning models that incorporate optimization functions to improve the ability of the model to identify specific security risks. The optimization function is defined as:

in this formula, x _n Representing the parameter set of the model, y represents the image data input into the model, and N is the number of model parameters. For example, a team is using a deep learning model to identify whether workers are not wearing helmets in a job site. In this case, the parameter set x of the model _n Including specific features in the image such as head shape, color, etc. Input image data y is Real-time video streaming from a job site. By substituting the data and the parameters into the optimization function, the model can calculate the gradient of the parameters, thereby adjusting and optimizing the model to improve the accuracy of identifying the safety risk.

By the method, the deep learning model can more accurately identify key features in the image and effectively distinguish whether potential safety hazards exist. For example, the model can accurately identify workers who are not wearing helmets and alert management teams in time. This not only improves the efficiency of security management, but also greatly reduces the potential risk of the construction site. Furthermore, by continually receiving new data from the job site, the model may continue to self-optimize and learn, thereby becoming more accurate and reliable over time.

S3, applying an integrated real-time analysis algorithm:

to ensure that potential security risks can be identified and alerted in time, teams employ the following integrated real-time analysis algorithm:

in this formula, t represents a time variable, and s represents a sensitivity factor of the data. The purpose of this function is to analyze the data collected from the cameras and sensors in real time and identify any anomalies or potential risks in time. For example, a project team is monitoring an area of a job site to identify if there are unauthorized personnel entering or if there are workers wearing safety helmets. In this case, the time variable t represents the length of time from the start of monitoring to the present, while the data sensitivity factor s is used to adjust the sensitivity of the alarm system. By inputting real-time data into this formula, the team can be alerted immediately when the data exhibits any abnormal pattern.

With such real-time analysis algorithms, the camera and sensor data is not only continuously monitored, but also an alarm can be raised immediately upon detection of any potential risk. In this way, a team can take measures quickly, such as dispatching security personnel to the scene, or notifying related personnel to take security measures. The method greatly improves the safety and response efficiency of the construction site, ensures the smooth progress of projects and ensures the safety of all staff. In summary, by combining deep learning, image processing technology and real-time analysis algorithm, project management team not only can accurately identify potential risk of construction site, but also can respond rapidly, thereby ensuring the safety and the high efficiency of construction progress.

Example 5: in a large office building construction project, a project management team decides to use machine vision algorithms developed by deep learning and image processing techniques, and data analysis techniques to comprehensively analyze data from a variety of sensors, including temperature, humidity and vibration, to detect risks including fire or structural instability.

First, the goal of a project team is to identify potential anomaly patterns from data collected from various sensors, such as early signs of fire or signs of structural instability. For this reason, the present embodiment employs an abnormality detection formula to analyze these data:

In this formula, x _i Represents the data points, μ, collected from the sensor _i Sum sigma _i Representing the mean and standard deviation of these data, respectively. For example, data collected by temperature sensors during the day shows an abnormal rise in temperature, which is an early sign of fire. Team calculates the mean (μ) of these temperature data _i ) And standard deviation (sigma) _i ) These data are then substituted into the anomaly detection formula. If the result of the formula exceeds a predetermined threshold, which will be seen as an indication of an abnormal pattern, the team will immediately take action, such as sending a security alarm or dispatching security personnel for field inspection. In addition, the team also analyzed data from humidity and vibration sensors. For example, if the vibration data suddenly changes, this indicates that there is a risk of instability of the building structure. By applying the same heteroOften, the team can identify these potential risks in time. This approach allows teams to monitor the safety of a job site from multiple angles and to conduct comprehensive risk assessment in conjunction with multiple types of sensor data. By the analysis method based on the multi-data fusion, a team can timely identify potential risks and take preventive measures to ensure the safety and stability of a construction site.

In order to further improve the accuracy and efficiency of risk prediction, the team uses the following risk identification model optimization formula:

G(p，q)＝∫ ₀ ^p (1-e ^-q·t )dt

in this formula, p represents the data set collected from the sensor and q represents the threshold for assessing risk. The purpose of this formula is to predict the probability of occurrence of a particular risk by calculating the risk accumulation value. For example, in analyzing data from a temperature sensor, a team finds that the temperature continues to rise over a period of time. The present embodiment takes the temperature data in this period as the data set p, and sets a temperature threshold q, for example, q=50 degrees celsius, as the early warning threshold for fire. By substituting these data into the risk identification model optimization formula, the team can calculate the probability of fire occurrence. This approach allows teams to monitor not only the current risk level, but also to predict the risk of future occurrence, thereby taking precautions ahead of time. For example, if the model predicts that the probability of fire risk exceeds a set safety threshold, the team may immediately take action, such as adding a spot patrol, preparing fire suppression equipment, or adjusting the construction plan to reduce the risk.

In summary, by combining deep learning, image processing technology and risk recognition model optimization formula, the project management team can realize comprehensive analysis of various sensor data including temperature, humidity, vibration and the like, and timely recognize and predict potential fire or structure instability risks. The application of the building construction management method based on the multi-data fusion not only improves the accuracy and the efficiency of risk management, but also provides strong support for ensuring the safety of construction sites and the smooth progress of projects.

Example 6: in a large office building construction project, in order to improve the safety management level and efficiency, a project management team takes a series of measures, utilizes an artificial intelligence algorithm to comprehensively analyze camera and sensor data, predicts and identifies various potential risks, and carries out real-time alarm when detecting the risks.

The step of implementing artificial intelligence algorithm to predict and identify potential risks:

first, camera and sensor data are analyzed using machine learning and deep learning algorithms

Project teams first employ machine learning and deep learning algorithms to analyze the vast amounts of data collected from cameras and various types of sensors. In order to extract useful features from these data, the present embodiment employs a feature extraction formula:

in this formula, x _i And y _i Representing features extracted from the image and sensor data, respectively. For example, x _i Representing the pixel intensity of a region in the image data, and y _i Representing the temperature values measured by the sensors in the same area.

In analyzing an area of a construction site, a team recognizes a person density (x _i ) And an abnormal increase in temperature of the region is measured by a temperature sensor (y _i ). By substituting these data into the feature extraction formula, the team can calculate a specific value indicating that the area is at risk for fire. By the method, the deep learning model can more accurately identify key features in the image and the sensor data, and effectively distinguish whether potential safety hazards exist. For example, the model can accurately identify abnormal temperature rise of a specific area, and can timely send fire alarms to a management team in combination with personnel density data. The method not only improves the accuracy and efficiency of safety monitoring, but also realizes more comprehensive risk assessment and pre-prediction on the construction site by integrating data from different sources And (5) measuring.

The project management team takes further measures: and implementing a real-time monitoring and early warning system.

The implementation steps of the real-time monitoring and early warning system are as follows:

second, continuous monitoring using real-time data analysis functions

In order to realize more effective real-time monitoring and timely early warning response, a team adopts a real-time data analysis function:

G(t)＝∫ ₀ ^t s·e ^-λs ds

in this formula, t represents time, s represents the signal intensity collected from the sensor, and λ is the attenuation coefficient for adjusting the sensitivity and response speed of data analysis. For example, teams monitor critical areas of the construction site, with particular concern for security risks, such as illegal intrusion or worker not wearing a helmet. Over a period of time (t), the data collected from the sensor of the present embodiment shows an abnormal activity pattern (e.g., s represents a sudden increase in activity intensity). By substituting these real-time data into the analysis function, the team can calculate the current risk level. If the result of the formula exceeds a preset threshold, this indicates that an emergency situation exists, such as an illegal intrusion or a security procedure not being followed. In this case, the early warning system will immediately activate, giving an alarm and informing the management team to take the corresponding action, such as sending a security personnel to the field for inspection or taking other security measures.

Through the real-time monitoring and early warning system, the project team can respond to the change condition of the construction site in real time, and the potential risks and dangers can be recognized and prevented in time. This not only increases the level of safety at the job site, but also enhances team control over potential problems.

Example 7: in a large office building construction project, a project management team decides to use environmental impact assessment tools in order to assess and minimize the impact of construction on the surrounding environment. The goal of this tool is to provide a comprehensive view to understand

S1, performing influence analysis by using environment influence model

First, teams employ environmental impact models to quantify the impact of construction activities on the surrounding environment. The formula of the model is defined as:

in this formula, x, y, z represent the measure of noise, dust and waste, respectively, while a, b, c are the impact weights of these environmental factors.

For example, a team would like to evaluate the impact of a particular construction phase on the environment, where the noise level (x) averages 70 db, the amount of dust (y) is 2 kg per hour, and the waste (z) is 50 cubic meters per day. These values are substituted into the model formula and the team can calculate the overall impact of noise, dust and waste on the environment at this stage.

By the method, teams can quantify the environmental impact of different construction stages and formulate corresponding mitigation strategies accordingly. For example, if the model shows that the noise level has a greater impact on surrounding communities, the team may take noise reduction measures, such as installing noise barriers or adjusting the working time. If the amount of dust is high, the in-situ watering frequency can be increased. For waste management, the team may optimize waste recovery and disposal flows to reduce environmental impact.

S2, application of a risk identification formula:

teams employ a risk identification formula to further analyze features extracted from the environmental data to predict environmental risk:

in this formula, d _i Representing features extracted from environmental data such as noise level, dust concentration or waste amount. The formula calculates the contribution of these features to the overall environmental risk and translates the data features into a risk assessment.

For example, team collects noise from job siteData of sound level, dust concentration and waste amount, and relevant characteristic values are extracted from these data by a specific analysis method. These characteristic values d _i Is input into the risk identification formula. By calculation, the formula can output an overall environmental risk assessment value. If the value exceeds a predetermined safety threshold, indicating a high environmental risk, corresponding mitigation measures need to be taken. By the method, the team can evaluate the immediate environment influence of the construction activity according to the actual measurement data and can predict the environment risk appearing in the future. This provides the basis for the present embodiment to take environmental protection measures in time, such as adjusting construction methods, optimizing waste treatment processes, or implementing noise reduction measures.

S3, evaluating function by using slow-down effect

To evaluate the effectiveness and persistence of the mitigation measures, the team uses the following formula:

in this formula, t represents time, u represents the effect of the mitigation measures, and α is the adjustment coefficient. The formula combines the rate of change of the slow effect over time (i.e. the rate of improvement or deterioration of the effect) with the influence of the adjustment coefficient on the effect. For example, a team implements a series of noise reduction measures to reduce the impact of construction noise on the surrounding environment. The present embodiment monitors the effect of these measures, i.e. the change in noise level (u), and records data over a period of time (t). By inputting these data into the mitigation effect evaluation function and setting the appropriate adjustment coefficients (α), the team can calculate the effect change of these measures over time. If the results of the formulas show that the effect of the mitigation measures increases gradually over time (i.e., du/dt is positive), these measures are indicated to be effective. Conversely, if the results show that the effect diminishes over time, readjustment or reinforcement is required. By the method, the team can take measures according to the environmental impact model and the risk identification result, and can monitor and adjust the effects of the measures in real time, so that the environmental impact of construction activities is effectively controlled.

S4, applying an adjustment feedback function:

to ensure that environmental impact is continually minimized and to maintain the utility of the measure, teams employ an adjustment feedback function:

in this formula, v represents an environmental influence variable such as noise level, dust concentration, or waste amount. Lambda is an adjustment factor that indicates the magnitude and speed of the adjustment required. The formula represents the rate of change of environmental impact variables over time and is an important tool for teams to guide environmental impact assessment and measure adjustment. For example, a team is monitoring the dust concentration (v) and finds that the dust concentration has risen for a certain time (t). The present embodiment inputs these data into the adjustment feedback function, and the rate of change is calculated. If this rate exceeds a set threshold, this indicates that an enhanced dust abatement is required. By the method, the team can monitor the environmental condition of the construction site in real time and timely adjust the slowing measures according to the feedback data. The method not only ensures more accurate and efficient environmental impact control, but also ensures the sustainability of construction activities and the satisfaction of communities.

In summary, the project management team can comprehensively evaluate the influence of the construction activity on the environment by combining multiple steps of the environmental influence evaluation tool, and implement effective slowing down and adjusting strategies. The environmental management based on the data and the scientific method not only improves the environmental protection performance of the construction project, but also ensures the realization of the social responsibility and sustainable development targets of the project.

The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made without departing from the spirit and scope of the invention, which is defined in the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims

1. The construction management method based on the multi-data fusion is characterized by comprising the following steps of:

firstly, a real-time multi-source data integration and analysis platform is adopted, which is used for integrating multi-source data of a building construction site, including material tracking, worker positions and machine states, into a unified platform, and processing information of various data sources from Internet of things equipment, unmanned aerial vehicle monitoring and personnel positioning systems through a data processing engine; secondly, adopting an intelligent prediction and risk assessment model, analyzing historical data and real-time data by utilizing machine learning and artificial intelligent technology to predict risk and construction progress, wherein the method comprises the steps of collecting the historical construction data to train the prediction model, and inputting new data in real time to perform risk assessment and progress prediction;

The method also comprises a self-adaptive resource optimization scheduling algorithm, wherein the allocation of resources including manpower, machinery and materials is automatically adjusted according to real-time data so as to optimize the construction efficiency, and the allocation of the resources is dynamically adjusted according to real-time scenes and predicted data; the intelligent safety monitoring system is combined, the safety condition of a construction site is monitored in real time by adopting a machine vision and sensing technology, and potential hazards are timely alerted, and the system comprises a high-resolution camera and a sensor which are deployed, and a software system for detecting potential safety threats is developed; and finally, adopting an environmental impact assessment tool for assessing and minimizing the impact of construction on the environment, including noise, dust and waste management, analyzing the impact of construction activities on the surrounding environment by creating a model, and formulating a slowing measure.

2. The construction management method based on multi-data fusion according to claim 1, wherein the real-time multi-source data integration and analysis platform comprises:

H(i,j,k)＝ln(i ^αk +e ^jβ )

3. The construction management method based on multi-data fusion according to claim 1, wherein the intelligent prediction and risk assessment model comprises the following steps:

I(r，s)＝∫e ^r·s dr

4. The construction management method based on multi-data fusion according to claim 1, wherein the adaptive resource optimization scheduling algorithm comprises the following steps: firstly, multidimensional data integration is carried out, and a data fusion function is applied:

G(a，b，t)＝a·e ^-bt

5. The building construction management method based on multi-data fusion according to claim 4, wherein the safety monitoring method implemented by combining machine vision and sensing technology is characterized in that: the method comprises the steps of utilizing a machine vision algorithm developed by deep learning and image processing technology to analyze image and video data captured by a high-resolution camera in real time so as to identify potential safety hazards including illegal invasion and unworn safety helmet;

6. The construction management method based on multi-data fusion according to claim 5, wherein the real-time analysis of the image and video data captured by the high-resolution camera uses:

s2, adopting a deep learning model to utilize an optimization function:

s3, adopting an integrated real-time analysis algorithm:

7. The construction management method based on multi-data fusion according to claim 5, wherein the machine vision algorithm developed by adopting deep learning and image processing technology is used for comprehensively analyzing various sensor data including temperature, humidity and vibration; by applying the anomaly detection formula:

and finally, adopting a risk identification model optimization formula:

8. The construction management method based on multi-data fusion according to claim 5, wherein the artificial intelligence algorithm predicts and identifies various potential risks comprises:

9. The method for managing building construction based on multi-data fusion according to claim 1, wherein the environmental impact assessment tool comprises:

s1, firstly, using an environmental impact model: