Civil engineering building monitoring system and method

Abstract

The invention belongs to the technical field of building monitoring, and discloses a civil engineering building monitoring system and method, and the system comprises a data collection module which collects the data, geological data and maintenance data of a civil engineering building; the environment data acquisition module acquires surrounding environment data of a building in real time; the underground water level monitoring module acquires underground water level data in real time; collecting meteorological data of the building area in real time; the building structure monitoring module collects building structure data in real time; the inspection module is used for monitoring the building in all directions; the image analysis module identifies abnormal conditions; the infrared data analysis module analyzes the infrared data; the comprehensive evaluation module comprehensively evaluates the abnormal condition; and the risk prediction module fuses the multi-source data and predicts potential problems and risks. According to the invention, abnormal conditions of the building are comprehensively evaluated; and the risk prediction module performs multi-source data fusion, comprehensively analyzes a large amount of different types of monitoring data, and predicts potential problems and risks. Preventive measures are taken in advance.

Description

Technical Field

[0001] The present application relates to the technical field of building monitoring, and more specifically, to a civil engineering building monitoring system and method. Background Art

[0002] Monitoring civil engineering structures enables real-time data collection on structural stress, strain, and displacement. This allows for the timely detection of structural anomalies during construction and operation, such as excessive deformation or crack expansion, thus providing a basis for structural safety assessment. This allows for the timely identification of potential safety hazards and proactive measures to address them, preventing serious accidents such as structural damage or collapse and safeguarding human life and property. For example, during bridge construction, monitoring the settlement and stress of piers allows for timely adjustments to prevent tilting or cracking if anomalies are detected. Civil engineering structure monitoring is extremely important.

[0003] The document with the prior art publication number CN115388853A provides a civil engineering construction monitoring system, including: a settlement monitoring tube, which is provided with an annular groove; a slip ring, which is sleeved on the annular groove on the settlement monitoring tube; an anchor rod assembly, which is slidably arranged horizontally on the slip ring; a horizontal cut-in assembly, which is arranged on the slip ring; the horizontal cut-in assembly includes a plurality of horizontal support plates. When the settlement monitoring tube is fully inserted into the ground, the anchor rod assembly and the horizontal cut-in assembly can expand outwards of the slip ring in succession, and be anchored into the soil by insertion and cutting respectively, so that the slip ring is combined with the underground soil layer. When the underground soil layer settles, it can drive the slip ring combined with it to settle together. By monitoring the settlement of the slip ring, the settlement of the underground soil layer can be monitored. This monitoring system does not require subsequent filling, so that the combination of the slip ring and the underground soil layer is more stable, thereby making the monitoring result more accurate.

[0004] Although the above-mentioned existing technical solutions can achieve relevant beneficial effects through the structure of existing technology, they still have the following defects: 1. Difficulty in detecting internal hidden dangers: For hidden dangers inside building structures, such as cracks, termite nests, rat holes, etc., existing technologies lack effective analysis methods. Traditional technologies often can only rely on manual sampling or simple physical detection methods, which are inefficient and have limited coverage. 2. Most current monitoring technologies rely only on a single data source to evaluate building abnormalities, and do not comprehensively consider the impact of multiple factors on buildings. A single data evaluation method cannot fully understand the true status of building abnormalities, which can easily lead to misjudgments on the nature, extent and scope of impact of abnormalities, thereby affecting the accuracy of subsequent maintenance decisions. 3. The lack of the ability to mine the association and potential patterns of multi-source data makes it impossible to accurately predict potential problems and risks in buildings.

[0005] In view of this, we propose a civil engineering construction monitoring system and method. Summary of the Invention

[0006] 1. Technical problems to be solved

[0007] The purpose of this application is to provide a civil engineering construction monitoring system and method that solves the technical problems raised in the above-mentioned background technology, realizes that the comprehensive assessment module combines the results of multiple monitoring modules to comprehensively evaluate abnormal conditions in the building; the risk prediction module performs multi-source data fusion and predicts potential problems and risks through comprehensive analysis of a large amount of different types of monitoring data. Multi-source data fusion can explore the correlation and potential patterns between different data, improve the accuracy and reliability of risk prediction; and predict potential risks in advance, so that managers can take preventive measures before risks occur.

[0008] 2. Technical solution

[0009] The technical solution of this application provides a civil engineering construction monitoring system, including:

[0010] Data collection module: collects civil engineering construction data, geological data, maintenance data, etc.; annotates the data as a reference sample;

[0011] Route planning module: plans the optimal route for drone flight based on the building structure and monitoring requirements;

[0012] Inspection module: including drones equipped with high-definition cameras and infrared scanners to conduct all-round monitoring of buildings;

[0013] Environmental data acquisition module: real-time collection of building surrounding environment data;

[0014] Groundwater level monitoring module: Groundwater level monitoring sensors are reasonably arranged around the building and near the foundation to obtain groundwater level data in real time.

[0015] Meteorological data acquisition module: real-time collection of meteorological data in the building area, including temperature, humidity, rainfall, wind speed and direction, etc.

[0016] Building structure monitoring module: multiple sensors (including crack detectors, strain gauges, levels, displacement meters, inclinometers, etc.) are rationally arranged at multiple locations on the building to collect building structure data in real time;

[0017] Image analysis module: pre-processes, extracts features, and analyzes and identifies collected images to promptly identify abnormalities;

[0018] Infrared data analysis module: Analyze infrared data and detect abnormal conditions (including cracks, termite nests, rat holes, etc.) in a timely manner;

[0019] Comprehensive assessment module: combines the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to comprehensively assess abnormal situations;

[0020] Risk prediction module: Integrates the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to perform multi-source data fusion and predict potential problems and risks;

[0021] Alarm module: includes an alarm, which will issue an alarm in time when abnormal conditions or potential risks are detected.

[0022] Interactive Module: Develop a user interface with comprehensive functionality and ease of use, implemented through web and mobile applications. Users can view real-time monitoring data on building settlement, tilt, deformation, and other indicators through the interface, displaying them in intuitive forms such as charts, curves, and 3D models.

[0023] Control center: and comprehensive assessment module: connect the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, risk prediction module, alarm module, building structure monitoring module and image analysis module to the network.

[0024] As an optional solution of the present invention, a comprehensive assessment module combines the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module, and the image analysis module to comprehensively assess abnormal conditions; the comprehensive assessment module includes the following steps:

[0025] 1. Data Collection: Collect data from each module. The environmental data acquisition module obtains environmental data around the building. The groundwater level monitoring module collects groundwater level data around the building and near the foundation using high-precision pressure or float-type water level gauges. The meteorological data acquisition module provides real-time meteorological data such as temperature, humidity, rainfall, and wind speed and direction in the building area. The building structure monitoring module transmits building structure data collected by various sensors deployed at different locations in the building. The image analysis module provides processed and analyzed image information; collects monitoring results from the infrared data analysis module;

[0026] 2. Data integration: Summarize and organize data from different modules, and associate them according to chronological order and dimensions such as building location to ensure data consistency and integrity, providing a comprehensive data foundation for subsequent comprehensive evaluation.

[0027] 3. Data preprocessing: clean and standardize the integrated data;

[0028] 4. Establish an evaluation index system: Based on civil engineering expertise and actual monitoring needs, identify key indicators for comprehensive evaluation. Weight each indicator based on its importance to the overall condition of the building, reflecting its relative importance in the comprehensive evaluation.

[0029] 5. Comprehensive evaluation and analysis:

[0030] Multi-factor correlation analysis: Preprocessed and standardized data is analyzed using a correlation analysis system based on an evaluation index system. This examines the interrelationships between different data types, such as the relationship between groundwater level rise and building foundation settlement, and the relationship between meteorological conditions (such as strong winds and heavy rain) and structural displacement. Through data analysis, potential patterns and trends are uncovered, and the combined impact of various factors on building conditions is determined.

[0031] Abnormality determination: Calculate and evaluate the integrated data based on the established evaluation indicator system and weight distribution. Set an evaluation threshold. When the comprehensive evaluation result exceeds the threshold, it is determined that an abnormality exists.

[0032] Determining the type and severity of anomalies: Further analyzing the specific indicators involved in the anomaly and combining professional knowledge to determine the type of anomaly, such as structural anomalies, environmental anomalies, or anomalies caused by meteorological factors. Furthermore, based on the degree to which each indicator deviates from the normal range, the severity of the anomaly is assessed, providing a basis for subsequent appropriate measures.

[0033] 6. Generate an Assessment Report: The comprehensive assessment results are compiled into a detailed report that includes the assessment timeframe, building locations involved, the type and extent of any anomalies found, and an analysis of key influencing factors. Relevant data charts are also included to make the report more intuitive and easier for civil engineering professionals and management personnel to understand. The assessment report is provided to relevant departments and personnel to provide a basis for decision-making regarding building maintenance, restoration, or further monitoring.

[0034] As an optional solution of the present invention, the risk prediction module integrates the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module, and the image analysis module to perform multi-source data fusion and predict potential problems and risks; the risk prediction module includes the following steps:

[0035] 1. Data Collection and Organization: Collect multi-source data, including environmental data, groundwater level data, and meteorological data. The infrared data analysis module identifies anomalies such as internal cracks, termite nests, and rat holes, along with corresponding temperature distribution and location information for abnormal areas. The image analysis module processes anomalies such as cracks, spalling, and deformation on the building surface, along with related image feature information. The building structure monitoring module also collects data collected by sensors such as crack meters, strain gauges, levels, displacement meters, and inclinometers deployed at various locations. This multi-source data is organized chronologically and by data type.

[0036] 2. Data preprocessing: preprocess the collected data, including data cleaning and data standardization;

[0037] 3. Time Series Analysis:

[0038] 3.1. Trend analysis: For the time series data of each monitoring indicator, such as groundwater level time series, structural displacement time series, etc., the moving average method, exponential smoothing method and other methods are used to analyze the long-term trend of the data.

[0039] 3.2. Seasonal analysis: Use seasonal decomposition methods (such as STL decomposition) to analyze indicators with seasonal changes.

[0040] 3.3 Periodicity Analysis: Use methods such as Fourier transform to identify periodic components in time series and determine the periodicity of indicator changes. For example, if traffic flow around a building is found to have a weekly periodic pattern, its potential impact on building vibration and other aspects can be assessed.

[0041] 3.4 Univariate Time Series Analysis: By building time series models, such as ARIMA and LSTM models, we predict indicator values ​​under normal conditions and compare them with actual observed values. When the deviation between the actual and predicted values ​​exceeds a certain threshold, we mark that period as a period of potential risk.

[0042] 3.5. Multivariate time series analysis: Considering the correlation between different indicators, Granger causality test and other methods are used to analyze the causal relationship between the indicators.

[0043] 4. Risk Prediction: Combining the results of univariate and multivariate time series analysis, a risk score is assigned to each time period and indicator. Scoring criteria are determined based on factors such as the degree of trend change, the frequency of outliers, and the strength of causal relationships between indicators. For example, time periods and indicators with dramatic trend changes, frequent outliers, and strong causal relationships with other key indicators receive higher risk scores. Based on historical data and engineering experience, a screening threshold for risk scores is set. Time periods and indicators with risk scores above the threshold are identified as potentially risky and serve as key focus points for subsequent finite element analysis.

[0044] 5. Finite Element Model Construction: Select an appropriate finite element model based on the building's structural type and complexity, as well as the potential risks identified through time-series analysis. For common building structures, such as frame structures and concrete structures, beam-column models and plate-shell models can be used. Using professional finite element analysis software, based on architectural design drawings and actual measurement data, accurately construct a 3D model of the building structure. Define the size, shape, material properties, and other parameters of each component to ensure the model accurately reflects the actual building structure.

[0045] Divide the grid appropriately based on the characteristics of the building structure and analysis requirements. For critical locations, such as beam-column joints and structural mutations, increase mesh density to improve calculation accuracy. For non-critical locations, appropriately reduce mesh density to reduce the amount of calculation. Accurately set the constitutive relationships, strength parameters, and other properties of building materials. Based on the actual foundation form and support conditions of the building, reasonably set boundary conditions, such as fixed supports, hinged supports, and elastic supports, to simulate the interaction between the building, the foundation, and the surrounding environment. At the same time, consider changes in boundary conditions under different working conditions, such as simulating the motion input of the foundation under earthquake conditions.

[0046] 6. Finite element analysis:

[0047] 6.1. Load application: Based on the potential risk periods and indicators determined by time series analysis and combined with the actual project conditions, the load conditions for finite element analysis are determined.

[0048] 6.2. Simulation Analysis: Run finite element analysis software to simulate the mechanical response of the building structure under different load conditions, including stress, strain, displacement, etc. Analyze the mechanical behavior of the structure under the influence of potential risk factors and explore the mechanical mechanisms that cause risks.

[0049] 6.3 Results Evaluation: Based on the finite element simulation results, assess the specific impact of potential risks on the building structure. Compare the mechanical performance indicators of the structure under normal operating conditions and potential risk conditions. Determine the severity of the risk and provide a basis for formulating appropriate risk response measures.

[0050] 7. Generate a risk report: Organize the results of time series analysis and finite element analysis to prepare a detailed risk report. The report includes information such as potential risk periods, indicators, the mechanical mechanisms of risk generation, specific impact levels, and risk severity classification. Relevant data charts and finite element simulation results are also included to make the report more intuitive and clear.

[0051] The present invention provides a civil engineering construction monitoring method, comprising the following steps:

[0052] S1. The data collection module collects civil engineering construction data, geological data, and maintenance data, etc.; it annotates the data and uses it as a reference sample;

[0053] S2. The environmental data acquisition module collects real-time environmental data about the building's surroundings; the groundwater level monitoring module obtains real-time groundwater level data. The meteorological data acquisition module collects real-time meteorological data about the building area, including temperature, humidity, rainfall, wind speed and direction, etc.

[0054] S3, building structure monitoring module rationally arranges multiple sensors (including crack detectors, strain gauges, levels, displacement meters and inclinometers, etc.) at multiple locations on the building to collect building structure data in real time;

[0055] S4, the route planning module plans the optimal route for the drone flight based on the building structure and monitoring requirements;

[0056] S5, the inspection module uses drones equipped with high-definition cameras and infrared scanners to conduct comprehensive monitoring of buildings along the planned route;

[0057] S6. The image analysis module performs preprocessing, feature extraction, and analysis and recognition on the collected images to identify abnormal situations in a timely manner;

[0058] S7, infrared data analysis module analyzes infrared data and promptly detects abnormal conditions (such as cracks, termite nests, rat holes, etc.);

[0059] S8, the comprehensive assessment module combines the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module and the image analysis module to comprehensively assess abnormal conditions;

[0060] S9, the risk prediction module integrates the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module and the image analysis module to perform multi-source data fusion and predict potential problems and risks;

[0061] S10. Alarm module: This includes an alarm that issues a warning when any abnormality or potential risk is detected. Users can view real-time monitoring data such as building settlement, tilt, and deformation through the interface.

[0062] 3. Beneficial effects

[0063] One or more technical solutions provided in the technical solution of this application have at least the following technical effects or advantages:

[0064] 1. The present invention uses an image analysis module to preprocess, extract features, and analyze and identify collected images, enabling timely identification of abnormalities such as cracks, spalling, and deformation on building surfaces. The infrared data analysis module analyzes infrared data to quickly identify hidden dangers such as cracks, termite nests, and rat holes within the building structure. This significantly shortens the time it takes to discover abnormalities and improves the accuracy of anomaly detection.

[0065] 2. The comprehensive assessment module combines the results of multiple monitoring modules to comprehensively evaluate any abnormalities in the building. This module considers the impact of multiple factors on the building, avoiding the one-sidedness of single-data assessments and enabling a more accurate assessment of the nature, extent, and scope of the abnormality.

[0066] 3. The risk prediction module integrates multi-source data, comprehensively analyzing large amounts of diverse monitoring data to predict potential issues and risks. Multi-source data fusion can uncover connections and underlying patterns between disparate data, improving the accuracy and reliability of risk predictions. Predicting potential risks in advance enables managers to take preventative measures before they occur, avoiding or mitigating losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is an overall schematic diagram of the civil engineering construction monitoring system disclosed in this application. DETAILED DESCRIPTION

[0068] The present application is further described in detail below with reference to the accompanying drawings.

[0069] Reference Figure 1 , an embodiment of the present application provides a civil engineering construction monitoring system, comprising:

[0070] Data collection module: collects civil engineering construction data (including drawings, construction data, etc.), geological data (including rock and soil stratification, rock and soil physical and mechanical properties (such as density, compression modulus, internal friction angle, etc.), groundwater level raw data, etc.) and maintenance data; annotates the data as a reference sample;

[0071] Route Planning: This module plans the optimal flight path for the drone based on the building's structure and monitoring requirements. Using the A* algorithm, the module takes into account the building's structural conditions, flight obstacles (such as building accessories and surrounding trees), altitude restrictions, and signal transmission range, to calculate the optimal path from the drone's starting point to each monitoring point. The algorithm incorporates real-time meteorological data, such as wind speed and direction, to dynamically adjust the flight path and ensure the drone's safety and stability.

[0072] Inspection module: including drones equipped with high-definition cameras and infrared scanners to conduct all-round monitoring of buildings;

[0073] Environmental Data Acquisition Module: This module collects real-time data about the building's surrounding environment. Vibration sensors are installed within a certain radius around the monitored building to capture vibration signals generated by construction activities such as piling. Furthermore, a total station or GPS is used to monitor displacement in surrounding construction areas, such as foundation pit excavation, which may affect soil stress distribution.

[0074] Groundwater level monitoring module: Groundwater level monitoring sensors are rationally arranged around the building and near the foundation, and high-precision pressure or float-type water level gauges are used to ensure accurate and real-time acquisition of groundwater level data.

[0075] Meteorological data acquisition module: real-time collection of meteorological data in the building area, including temperature, humidity, rainfall, wind speed and direction, etc.

[0076] Temperature Monitoring: Temperature sensors, such as thermistors, should be evenly distributed across the building's different orientations, heights, and key structural areas. These sensors should possess excellent accuracy and stability, adapting to varying ambient temperature ranges. Temperature sensors should be connected to tilt and deformation sensors through the same data acquisition module to ensure synchronized data collection.

[0077] Wind speed and direction monitoring: Install anemometers and rain gauges on the top of the building and in the surrounding open areas. The anemometer can be a three-cup type or ultrasonic type, and the rain gauge can be a tipping bucket type or siphon type.

[0078] Building structure monitoring module: multiple sensors (including crack detectors, strain gauges, levels, displacement meters, inclinometers, etc.) are rationally arranged at multiple locations on the building to collect building structure data in real time;

[0079] For frame structures, strain gauges and displacement meters are installed at key nodes and along the column shaft to monitor column stress, strain, and displacement. For masonry structures, crack observation instruments are placed at vulnerable areas of the wall, and settlement observation points are set up at the foundation, using a level or static level for settlement monitoring. These monitoring devices are strategically distributed based on the building's structural characteristics and connected to a data acquisition system.

[0080] For concrete structures, strain gauges and temperature-compensating strain gauges are embedded in various locations to monitor strain changes in the concrete. Humidity sensors are also used to monitor internal humidity in the concrete for creep and shrinkage analysis. For steel structures, stress sensors are installed on key load-bearing components. The corrosion potential of the steel is regularly monitored and, combined with stress data, the impact of corrosion on the structure is assessed. This sensor data is processed by conditioning circuits and transmitted to a data acquisition terminal.

[0081] Image Analysis Module: This module preprocesses, extracts features, and analyzes and identifies acquired images to promptly identify anomalies. It also builds an anomaly recognition model based on machine learning or deep learning. By training the model on a large number of normal and abnormal building images, it can accurately identify anomalies such as cracks, spalling, and deformation on building surfaces. During model training, the model parameters are continuously optimized to improve recognition accuracy and generalization capabilities.

[0082] Infrared data analysis module: Analyze infrared data and detect abnormalities (such as cracks, termite nests, rat holes, etc.) in a timely manner;

[0083] Comprehensive assessment module: combines the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to comprehensively assess abnormal situations;

[0084] Risk prediction module: Integrates the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to perform multi-source data fusion and predict potential problems and risks;

[0085] Alarm module: includes an alarm, which will issue an alarm in time when abnormal conditions or potential risks are detected.

[0086] Interactive Module: Develop a comprehensive, user-friendly interactive interface, accessible through web and mobile applications. Users can view real-time monitoring data on building settlement, tilt, and deformation, displayed in intuitive formats such as charts, curves, and 3D models. A historical data query function facilitates analysis of structural status trends.

[0087] Control center: and comprehensive assessment module: connect the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, risk prediction module, alarm module, building structure monitoring module and image analysis module to the network.

[0088] Furthermore, the route planning module plans the optimal route for the drone flight based on the building structure and monitoring requirements; the steps include:

[0089] 1. Collect Basic Information: Comprehensively gather structural information about the building, including detailed data such as its layout, height, and the location and dimensions of its ancillary facilities. Clarify monitoring requirements and determine the specific locations and requirements for each monitoring point. Understand obstacles within the flight area, such as the location and height of surrounding trees and the distribution of building ancillary facilities. Understand parameters related to flight altitude restrictions and signal transmission range.

[0090] 2. Initialize the A* algorithm: Determine the starting point of the drone and the coordinates of each monitoring point in the algorithm. Set the parameters required by the A* algorithm, such as the heuristic function (e.g., Manhattan distance or Euclidean distance).

[0091] 3. Build an environmental model: Based on the collected information, the algorithm constructs a virtual environment model that includes obstacles, flight altitude restrictions, and signal transmission range. Obstacles are accurately represented in the model so that they can be recognized and processed by the algorithm.

[0092] 4. Calculate the initial optimal path: Using the A* algorithm, we calculate the initial optimal path for the drone from its starting point to each monitoring point within the constructed environmental model. The algorithm searches for a path with the lowest cost based on the specified heuristic function and environmental model.

[0093] 5. Dynamic Path Adjustment: Real-time meteorological data, such as wind speed and direction, is collected within the flight area. This data is integrated into the A* algorithm, allowing it to account for the impact of wind speed and direction on the drone's flight path. Based on real-time meteorological data, the algorithm assesses the impact of wind direction on the initial optimal path. If wind direction threatens to destabilize the drone's flight or deviate from the optimal path, the algorithm re-searches the path and dynamically adjusts the drone's flight path. This path adjustment adheres to the principles of the A* algorithm, ensuring that the new path is the optimal choice after accounting for meteorological factors.

[0094] 6. Path Verification and Optimization: Verify the adjusted path to see if it meets requirements such as flight altitude restrictions and signal transmission range. If there are any issues with the path, further optimize the path until all conditions are met.

[0095] 7. Output Optimal Path: The optimal path from the starting point to each monitoring point is output to the inspection module. The path information can be output as a coordinate sequence or other format suitable for UAV flight control.

[0096] Furthermore, the image analysis module pre-processes, extracts features, and analyzes and identifies the collected images to promptly identify abnormal situations. This includes the following steps:

[0097] 1. Image Acquisition: High-definition cameras mounted on drones in the inspection module capture images of civil engineering structures, ensuring coverage of all key building components and capturing images under varying lighting conditions, angles, and other conditions to improve the generalization capabilities of subsequent models. The captured images are annotated to distinguish between normal building images and those containing abnormalities such as cracks, spalling, and deformation. For abnormal images, the location and type of the abnormal areas are precisely annotated to provide accurate label information for subsequent model training. The annotated image dataset is divided into a specific proportion. The training set is used for model learning, the validation set is used to adjust model parameters and prevent overfitting during training, and the test set is used to ultimately evaluate model performance.

[0098] 2. Image preprocessing: Preprocess the collected images, including normalization, image enhancement, grayscale, and noise reduction; use filtering algorithms (such as Gaussian filtering, median filtering, etc.) to remove noise interference in the image, making the image clearer and facilitating subsequent feature extraction and analysis.

[0099] 3. Feature extraction: Feature extraction is performed on the preprocessed image. The extracted features include color, texture and shape. Feature extraction algorithms such as SIFT (Scale Invariant Feature Transform), SURF (Speeded Robust Features), HOG (Histogram of Oriented Gradients), etc. are used to extract local features and texture features of the image.

[0100] 4. Model Construction: Based on the task requirements and data characteristics, select a convolutional neural network machine learning model and build the architecture of the anomaly recognition model. In the convolutional neural network (CNN) model, add an appropriate number of convolutional layers, pooling layers, and fully connected layers, and determine parameters such as the number of neurons and activation functions in each layer. In the model's output layer, set the appropriate output nodes based on the number of anomaly categories and use an appropriate activation function (such as the Softmax function) for classification.

[0101] Use the training set to train the model, continuously adjusting model parameters to minimize the model's loss function (such as the cross-entropy loss function). Use the validation set to evaluate the model and adjust parameters to prevent overfitting. Optimization algorithms (such as stochastic gradient descent (SGD) and Adam) can be used to accelerate model convergence. Model fusion, data augmentation, and regularization (such as L1 and L2 regularization) can be used to further optimize model performance and improve recognition accuracy and generalization. Model hyperparameters (such as the learning rate, batch size, and number of iterations) are continuously adjusted, using grid search and random search to find the optimal hyperparameter combination. Use the test set to perform a final evaluation of the trained model, calculating metrics such as accuracy, recall, and F1 score to assess model performance and generalization. Based on the results of the model evaluation, the model can be continuously optimized. Collect more image data for annotation and training, and further adjust the model's parameters and architecture to improve recognition accuracy and stability.

[0102] 5. Analysis and Identification: The preprocessed and feature-extracted images are fed into the trained model. The model then classifies the images based on the learned features and patterns, determining whether anomalies exist and the type of anomaly. For identified anomalies, the model further analyzes the specific information about the abnormal area, such as the length and width of the cracks and the area of ​​spalling.

[0103] 6. Result output: The identified abnormalities are output in an intuitive manner, such as generating reports, marking images of abnormal areas, etc., to facilitate relevant personnel to review and analyze.

[0104] Furthermore, the infrared data analysis module analyzes the infrared data to promptly detect abnormalities (such as cracks, termite nests, rat holes, etc.); the following steps are included:

[0105] 1. Data preprocessing: preprocess the collected infrared data, including format conversion, denoising and temperature calibration;

[0106] Format conversion: Convert infrared data into common formats, such as common image formats (.bmp, .jpg, etc.) or professional infrared data formats (.hdr, etc.), to facilitate processing using various analysis tools.

[0107] Denoising: Filtering algorithms are used to remove noise from infrared data. For example, a median filter is used to sort the values ​​of each pixel and its neighbors, taking the median value as the new value for that pixel, effectively removing salt and pepper noise. For Gaussian noise, a Gaussian filter is used to smooth the data by taking a weighted average of the neighboring pixel values, making the infrared image clearer.

[0108] Temperature Calibration: Due to environmental factors (such as temperature and humidity) and device errors, the collected temperature data may be biased. To ensure accurate temperature data, calibrate the raw temperature data based on the infrared scanner calibration parameters and environmental monitoring data (such as real-time temperature and humidity provided by the meteorological data acquisition module).

[0109] 2. Feature extraction: including temperature feature extraction and shape and texture feature extraction;

[0110] Temperature feature extraction: Calculate the mean and standard deviation of temperature. Divide the calibrated infrared image into multiple small areas and calculate the mean and standard deviation of the temperature within each area. These statistical features reflect the concentration trend and dispersion of temperature within the area, helping to identify areas of temperature anomalies.

[0111] Calculate the temperature gradient for each pixel in the image to detect areas of drastic temperature changes, which may correspond to anomalies in the building structure, such as changes in heat transfer through cracks.

[0112] Shape and texture feature extraction: Use edge detection algorithms (such as the Canny algorithm) to process infrared images and extract edge information of the temperature distribution. These edges may correspond to the boundaries of building structures or the outlines of abnormal areas. The Canny algorithm first performs Gaussian filtering on the image for smoothing, then calculates the image gradient amplitude and direction, then performs non-maximum suppression on the gradient amplitude, and finally obtains the final edge image through double threshold detection and edge connection. Analyze the texture features of infrared images using methods such as the gray-level co-occurrence matrix (GLCM). GLCM can describe the probability of different grayscale value pairs appearing in a certain spatial position relationship in an image. By calculating the statistics of GLCM (such as contrast, correlation, energy, and entropy), image texture information is obtained to distinguish the differences in texture features between normal and abnormal areas.

[0113] 3. Abnormality Identification: Set temperature thresholds based on the normal temperature range of building materials and historical data. When the temperature mean or gradient amplitude in a certain area exceeds the set threshold, the area is marked as a possible abnormal area. For example, for concrete structures, the normal temperature is within a certain range. If the temperature in a certain area exceeds the upper limit of the range by a certain degree, there may be an abnormality, such as an internal heat source (such as heat generated by microbial activity in termite nests) or a structural defect (such as cracks causing abnormal heat loss).

[0114] 4. Model Construction: Select a support vector machine (SVM) model and collect and annotate a large amount of infrared image data containing both normal and abnormal conditions (such as cracks, termite nests, and rat holes). Use this annotated data to train a machine learning model. By finding the optimal classification hyperplane, the feature vectors of normal and abnormal infrared images are classified into different categories. During training, an appropriate kernel function (such as the radial basis kernel function) and penalty parameters are selected to improve the model's classification performance. The feature-extracted infrared image data is input into the trained model. The model classifies the images based on the learned feature patterns, determining the presence and type of anomalies.

[0115] 5. Result Verification: Cross-validation is used to further verify the accuracy of identified abnormal results. The identified abnormal areas are compared with known abnormal samples or expert annotation results, and evaluation indicators such as accuracy and recall are calculated to evaluate model performance. If the accuracy is low, it is necessary to readjust the model parameters or increase the training data, and retrain and identify. For areas determined to be abnormal, the image spatial coordinate information is used in combination with the building structure drawings to accurately determine the location of the abnormality in the building. For example, matching the coordinates of the abnormal area in the infrared image with the coordinates of the building's three-dimensional model can accurately point out the specific location of cracks, termite nests, or rat holes in the building structure, providing accurate information for subsequent maintenance and treatment.

[0116] 6. Result Output: Infrared data analysis results are compiled into a detailed report, including the type, location, severity assessment (e.g., crack length and width, approximate range of termite nests), and related analysis charts (e.g., temperature distribution maps, and abnormal area markers). Visualization software or platforms provide intuitive displays of infrared images and marked abnormal areas. Abnormal areas can be highlighted using different colors or markers, allowing relevant personnel to quickly understand the building's infrared characteristics and abnormal distribution. Interactive features are also provided, allowing users to view detailed temperature data and analysis results for specific areas, facilitating further analysis and decision-making.

[0117] Furthermore, the comprehensive assessment module combines the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module, and the image analysis module to comprehensively assess abnormal conditions; the comprehensive assessment module includes the following steps:

[0118] 1. Data collection: Collect data from each module, and obtain environmental data around the building from the environmental data acquisition module, such as vibration signals generated by construction activities such as piling captured by vibration sensors, and displacement monitoring data of surrounding construction areas such as foundation pit excavation using total stations or satellite positioning systems (GPS). The groundwater level monitoring module collects groundwater level data around the building and near the foundation using high-precision pressure or float-type water level gauges. The meteorological data acquisition module provides real-time meteorological data such as temperature, humidity, rainfall, wind speed and direction in the building area. The building structure monitoring module receives building structure data collected by various sensors (such as crack meters, strain gauges, levels, displacement meters and inclinometers, etc.) deployed at different locations in the building, including various types of monitoring data corresponding to different structural types (frame structures, masonry structures, concrete structures, steel structures, etc.). The image analysis module provides processed and analyzed image information, such as identified abnormal conditions such as cracks, spalling, and deformation on the building surface. Collect the monitoring results of the infrared data analysis module;

[0119] 2. Data integration: The data from different modules are aggregated and organized, and correlated according to chronological order and building location to ensure data consistency and integrity, providing a comprehensive data foundation for subsequent comprehensive evaluation.

[0120] 3. Data preprocessing: clean and standardize the integrated data;

[0121] Data cleaning: Check the collected data for missing values, outliers, and other issues. For missing values, interpolation methods (such as linear interpolation and polynomial interpolation), mean filling, and adjacent value filling are used to supplement them, depending on the data characteristics and actual situation. Outliers are identified, corrected, or eliminated by setting a reasonable data range and conducting statistical analysis (such as the 3σ principle) to ensure data quality.

[0122] Standardization: Since the dimensions and numerical ranges of data in different modules may vary, the data is standardized to facilitate comprehensive analysis. All types of data are uniformly converted to the same numerical range.

[0123] 4. Establish an evaluation indicator system: Based on professional knowledge of civil engineering construction and actual monitoring needs, identify key indicators for comprehensive evaluation. For building structural safety, indicators include structural displacement, crack width, and strain change. For environmental impacts, indicators include vibration intensity caused by surrounding construction activities and soil displacement caused by foundation pit excavation. For meteorological factors, consider the impact of wind on the building (by comparing wind force levels with building wind resistance design standards) and the magnitude of groundwater level changes caused by rainfall. Assign indicator weights and use methods such as the Analytic Hierarchy Process (AHP) and expert scoring to determine the weight of each evaluation indicator. Different weights are assigned based on the importance of each indicator to the overall building condition, reflecting its relative importance in the comprehensive evaluation. For example, the displacement indicator for key parts of the building structure may be weighted relatively high, while the indicators for some minor environmental factors may be weighted relatively low.

[0124] 5. Comprehensive evaluation and analysis:

[0125] Multi-factor correlation analysis: Preprocessed and standardized data is analyzed based on a correlation analysis system. The correlation between different types of data is observed, such as the relationship between groundwater level rise and building foundation settlement, and the relationship between meteorological conditions (such as strong winds and heavy rain) and changes in building structure displacement. Through data analysis, potential patterns and trends are discovered to determine the combined impact of various factors on the building's condition. The comprehensive evaluation model is:

[0126] E=[Σ n i＝1 (w i (t)x i )]+[Σ(β ij w i (t)w j (t)x i x j )]; 1≤i≤j≤n;

[0127] w i (t) = w i (0)*[1+γ i *△x i (t) / (△ — x i )]*a i (t);

[0128]

[0129] Where E is the comprehensive evaluation value. E is equal to the pre-set evaluation threshold E thresh In contrast, if E>E thresh , it is determined that the building has an abnormal condition, providing a key basis for building condition assessment. i(t) is the weight of the i-th evaluation indicator at time t, reflecting the importance of the indicator at time t; x i is the value of the ith evaluation index after normalization. j is the jth evaluation index value after normalization. ij is the coefficient reflecting the interaction strength between the i-th indicator and the j-th indicator, and its value range is [-1,1]; w i (t) and w j (t) are the weights of the i-th and j-th evaluation indicators at time t; x i and x j is the corresponding standardized evaluation index value. i (0) is the initial weight, which is the importance of the indicator at the initial setting; γ i is the adjustment coefficient related to the i-th indicator, which determines the sensitivity of the indicator weight to the fluctuation of monitoring data; △x i (t) is the difference between the value of the ith indicator at time t and the mean value of the indicator in the recent period, reflecting the degree of deviation of the indicator data at the current moment from the recent average level; — x i is △x in the recent period i The average fluctuation value of a i (t) is a weight correction function based on the characteristics of the building at different stages of use, which modifies the weights considering the characteristics of the building at different stages. 1i is an adjustment parameter related to indicator i, which mainly affects the transition from the new construction period to the stable use period of the building. This parameter determines the growth rate of the weight of indicator i over time during the new construction period. For example, during the new construction period, as the building gradually comes into use, the importance of some key structural indicators (such as key node strain, important part displacement, etc.) to the overall structural safety assessment may gradually increase. 1i is used to quantify the growth of this importance. 2i This is an adjustment parameter related to indicator i, primarily used during the transition from a building's stable use period to its aging period. Once a building enters the aging period, the performance of its structural materials gradually degrades, increasing structural safety risks. The importance of some structural safety-related indicators to the overall assessment increases. t represents the time variable. T is the building's designed service life, a fixed value determined by factors such as its type, intended use, and design standards. T1 is the demarcation point between the building's new construction period and its stable use period. This time value is determined based on the building's actual conditions and is generally reached some time after completion, when interior decoration, equipment commissioning, and other work are essentially complete, and the structure enters a relatively stable state of use. T2 is the demarcation point between the building's stable use period and its aging period. This value is determined based on factors such as the durability of the building's structural materials and the building's environment.

[0130] Abnormality Assessment: Calculate and evaluate the integrated data based on an established evaluation index system and weighted distribution. Set an assessment threshold, and when the comprehensive assessment results exceed the threshold, an abnormality is determined. For example, if the combined assessment values ​​of multiple indicators, such as structural displacement, crack width, and ambient vibration, exceed a pre-defined safety range, the building is considered abnormal.

[0131] Determining the type and severity of anomalies: Further analysis of the specific indicators involved in the anomaly is conducted, and combined with professional knowledge, the type of anomaly, such as structural anomaly, environmental anomaly, or anomaly caused by meteorological factors, is determined. Furthermore, the severity of the anomaly is assessed based on the degree to which each indicator deviates from the normal range, providing a basis for subsequent appropriate measures. For example, crack width and structural displacement can be used to determine whether the structural anomaly is mild, moderate, or severe.

[0132] 6. Generate an Assessment Report: The results of the comprehensive assessment are compiled into a detailed report, including the assessment timeframe, building locations involved, the type and extent of any anomalies found, and an analysis of key influencing factors. Relevant data charts (such as trend graphs showing changes in various indicators over time and the location and distribution of anomalies within the building structure) are also included to make the report more intuitive and easier for civil engineering professionals and relevant management personnel to understand. The assessment report is fed back to relevant departments and personnel to provide a basis for decision-making regarding building maintenance, restoration, or further monitoring.

[0133] Furthermore, the risk prediction module integrates the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module, and the image analysis module to perform multi-source data fusion and predict potential problems and risks. The steps include:

[0134] 1. Data Collection and Collation: Collect multi-source data, including environmental data, groundwater level data, and meteorological data. The infrared data analysis module identifies anomalies such as internal cracks, termite nests, and rat holes, along with corresponding temperature distribution and location information for abnormal areas. The image analysis module processes anomalies such as cracks, spalling, and deformation on the building surface, along with related image feature information such as crack length, width, and spalling area. The building structure monitoring module aggregates data collected by sensors such as crack meters, strain gauges, levels, displacement meters, and inclinometers deployed at various locations. This data covers monitoring data for various structural types under both routine operation and special operating conditions. This multi-source data is organized chronologically and by data type, and stored in a unified database to ensure data integrity and accuracy.

[0135] 2. Data preprocessing: preprocess the collected data, including data cleaning and data standardization;

[0136] 3. Time Series Analysis:

[0137] 3.1 Trend Analysis: For each monitoring indicator's time series data, such as groundwater level time series and structural displacement time series, we use methods such as moving average and exponential smoothing to analyze the long-term trend of the data. For example, we can use the moving average method to calculate the average trend of groundwater level changes over the past few months to determine whether the groundwater level is rising, falling, or remaining stable.

[0138] 3.2 Seasonal Analysis: Use seasonal decomposition methods (such as STL decomposition) to analyze indicators with seasonal variations, such as temperature and rainfall in meteorological data. Decompose the time series into trend terms, seasonal terms, and residual terms to determine the impact of seasonal factors on the indicators. For example, analysis may reveal that high summer rainfall may affect building foundations.

[0139] 3.3 Periodicity Analysis: Use methods such as Fourier transform to identify periodic components in time series and determine the periodicity of indicator changes. For example, if traffic flow around a building is found to have a weekly periodic pattern, its potential impact on building vibration and other aspects can be assessed.

[0140] 3.4 Univariate Time Series Analysis: By building time series models, such as ARIMA and LSTM models, we predict indicator values ​​under normal conditions and compare them with actual observed values. When the deviation between the actual and predicted values ​​exceeds a certain threshold, we mark that period as potentially risky. For example, when using an ARIMA model to predict structural strain, if the actual strain exceeds the predicted range, we preliminarily determine that the structure may be at risk during that period.

[0141] 3.5 Multivariate Time Series Analysis: Considering the correlations between different indicators, methods such as Granger Causality Tests are used to analyze the causal relationships between them. For example, this can be used to test whether rising groundwater levels are the cause of changes in building foundation displacement. By analyzing multivariate time series data, combinations of indicators with interrelated and unusual changes can be identified, thereby determining which sets of indicators may pose risks.

[0142] 4. Risk Prediction: Combining the results of univariate and multivariate time series analysis, a risk score is assigned to each time period and indicator. Scoring criteria are determined based on factors such as the degree of trend change, the frequency of outliers, and the strength of causal relationships between indicators. For example, time periods and indicators with dramatic trend changes, frequent outliers, and strong causal relationships with other key indicators receive higher risk scores. Based on historical data and engineering experience, a screening threshold for risk scores is set. Time periods and indicators with risk scores above the threshold are identified as potentially risky and serve as key focus points for subsequent finite element analysis.

[0143] R(t,i)=w1(t,i)S tr (t,i)+w2(t,i)f 异 (t,i)+w3(t,i)S 因 (t,i)+S 交互 ;

[0144] S 交互 =Σ[λ lm (t,i)S 因 (t,il)S 因 (t,i m )];1≤l<m≤m;

[0145] S 因 (t,i)=∑ m l＝1 F ijl ;w1(t,i)+w2(t,i)+w3(t,i)=1;

[0146] w1(t,i)=a1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+a2(t,i)f 异 (t,i)+a3(t,i)S 因

[0147] (t,i)];

[0148] W2(t,i)=B1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+B2(t,i)f 异 (t,i)+B3(t,i)S 因

[0149] (t,i)];B1(t,i)+B2(t,i)+B3(t,i)=1;

[0150] W3(t,i)=C1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+C2(t,i)f 异 (t,i)+C3(t,i)S 因

[0151] (t,i)]; C1(t,i)+C2(t,i)+C3(t,i)=1; where R(t,i) is the risk score function value, which is used to quantify the risk level corresponding to indicator i in time period t. t represents the time period. i represents the indicator, which refers to various monitoring indicators obtained from multi-source data collection, such as groundwater level, structural displacement, temperature, humidity, etc. Each indicator reflects the status information of the building from a different perspective. By analyzing and evaluating different indicators, we can fully understand the status of the building. w1(t,i) reflects the trend change degree score S tr (t,i) in the risk score; w2(t,i) reflects the frequency of outliers f 异 (t,i); w3(t,i) represents the causal relationship strength score S 因 The weight of (t,i). S tr (t,i) is the trend change score. The trend change of indicator i is measured by calculating the slope of the time series in period t. 异 (t,i) is the frequency of outliers, that is, the ratio of the number of outliers for indicator i to the total data volume during time period t. Outliers are data points that deviate significantly from the normal data distribution and are identified through certain statistical methods (such as box plot analysis). It is used to quantify the impact of the frequency of abnormalities in indicator i on risk. A higher frequency of outliers may indicate instability or abnormal conditions in the building structure or related factors, thereby increasing the risk of the building. 因 (t,i) is the causal relationship strength score. The F value of the Granger causality test measures the causal relationship strength between indicator i and other indicators with causal relationships. Other indicators with causal relationships with indicator i are j1, j2, …, j m , F ijl is the F value of the Granger causality test between indicator i and jl. It reflects the closeness of the causal relationship between indicator i and other related indicators, and is used to reflect the impact of the causal relationship between indicator i and other indicators on building risk. lm (t,i) reflects the coefficient of the interaction strength between the causal relationship between indicator il and indicator im. Its value range is determined according to the actual situation. It is a quantitative parameter to measure the degree of synergy or antagonism between different causal relationships. By reasonably determining λ lmThe value of (t,i) (which can be calculated through data analysis, expert experience, or machine learning) can more accurately reflect the complex connections between different causal relationships in building risks, improving the accuracy and reliability of risk scoring. α1(t,i) is a coefficient specifically used to adjust the weight of the trend change score in w1(t,i) based on the current building status and historical data changes. Its value changes dynamically and is closely related to the trend characteristics of indicator i in the building during time period t. a2(t,i) is a coefficient used to adjust the relative weight of the frequency of outliers in the calculation of w1(t,i). It is also dynamically calculated based on the current building status and historical data changes, and is related to the degree to which outliers for indicator i in time period t affect the trend change weight adjustment. a3(t,i) is a coefficient used to adjust the relative weight of the causal relationship strength score between indicators. B1(t,i) is a coefficient used to adjust the relative weight of the trend change score in the calculation of w2(t,i) based on the current building status and historical data changes. B2(t,i) is a coefficient used to adjust the weight of the frequency of outliers in w2(t,i). B3(t,i) is a coefficient used to adjust the relative weight of the causal relationship strength score between indicators. C1(t,i) is a coefficient used to adjust the relative weight of the trend change score in the calculation of w3(t,i) based on the current state of the building and changes in historical data. C2(t,i) is a coefficient used to adjust the relative weight of the frequency of outliers in the calculation of w3(t,i). C3(t,i) is a coefficient used to adjust the weight of the causal relationship strength score in w3(t,i). Based on the dynamic changes in the current state of the building and historical data, it is closely related to the causal relationship characteristics of indicator i in time period t.

[0152] 5. Finite Element Model Construction: Select an appropriate finite element model based on the building's structural type and complexity, as well as the potential risks identified through time-series analysis. For common building structures, such as frame structures and concrete structures, beam-column models and plate-shell models can be used. Using professional finite element analysis software (such as ANSYS and ABAQUS), based on architectural design drawings and actual measurement data, accurately construct a three-dimensional model of the building structure. Define the size, shape, material properties, and other parameters of each component to ensure the model accurately reflects the actual building structure.

[0153] Divide the grid appropriately based on the characteristics of the building structure and analysis requirements. For critical locations, such as beam-column joints and structural mutations, increase mesh density to improve calculation accuracy. For non-critical locations, appropriately reduce mesh density to reduce the amount of calculation. Accurately set the constitutive relationships, strength parameters, and other properties of building materials. Based on the actual foundation form and support conditions of the building, reasonably set boundary conditions, such as fixed supports, hinged supports, and elastic supports, to simulate the interaction between the building, the foundation, and the surrounding environment. At the same time, consider changes in boundary conditions under different working conditions, such as simulating the motion input of the foundation under earthquake conditions.

[0154] 6. Finite element analysis:

[0155] 6.1. Load Application: Based on the potential risk periods and indicators identified through time series analysis and in combination with the actual project conditions, the load conditions for finite element analysis are determined. For example, for potential risks caused by rising groundwater levels, corresponding water pressure loads are applied; for strong winds, wind loads of varying directions and intensities are applied based on meteorological data.

[0156] 6.2. Simulation Analysis: Run finite element analysis software to simulate the mechanical response of the building structure under different load conditions, including stress, strain, and displacement. Analyze the mechanical behavior of the structure under the influence of potential risk factors and explore the mechanical mechanisms that cause risks. For example, by simulating the effect of rising groundwater levels on the foundation, analyze the changes in foundation stress distribution and determine the location and form of potential foundation damage.

[0157] 6.3 Results Evaluation: Based on the finite element simulation results, assess the specific impact of potential risks on the building structure. Compare the mechanical performance indicators of the structure under normal operating conditions and potential risk conditions, such as whether stress exceeds the material allowable stress, whether displacement is within the allowable range, and whether the overall stability of the structure is affected. Determine the severity of the risk and provide a basis for formulating appropriate risk response measures.

[0158] 7. Generate risk report: Organize the results of time series analysis and finite element analysis and prepare a detailed risk report. The report content includes information such as potential risk periods, indicators, mechanical mechanisms of risk generation, specific impact levels, and risk severity levels. At the same time, attach relevant data charts and finite element simulation result charts to make the report more intuitive and clear. Based on the risk assessment results, put forward targeted risk response suggestions. For high-risk situations, it is recommended to immediately take measures such as structural reinforcement and foundation treatment; for medium-risk situations, increase monitoring frequency and formulate emergency plans; for low-risk situations, continue to pay attention to changes in indicators and conduct regular reviews. Provide decision support for the safety management and maintenance of buildings.

[0159] The present invention provides a civil engineering construction monitoring method, comprising the following steps:

[0160] S1. The data collection module collects civil engineering construction data, geological data, and maintenance data, etc.; it annotates the data and uses it as a reference sample;

[0161] S2. The environmental data acquisition module collects real-time environmental data about the building's surroundings; the groundwater level monitoring module obtains real-time groundwater level data. The meteorological data acquisition module collects real-time meteorological data about the building area, including temperature, humidity, rainfall, wind speed and direction, etc.

[0162] S3, building structure monitoring module rationally arranges multiple sensors (including crack detectors, strain gauges, levels, displacement meters and inclinometers, etc.) at multiple locations on the building to collect building structure data in real time;

[0163] S4, the route planning module plans the optimal route for the drone flight based on the building structure and monitoring requirements;

[0164] S5, the inspection module uses drones equipped with high-definition cameras and infrared scanners to conduct comprehensive monitoring of buildings along the planned route;

[0165] S6. The image analysis module performs preprocessing, feature extraction, and analysis and recognition on the collected images to identify abnormal situations in a timely manner;

[0166] S7, infrared data analysis module analyzes infrared data and promptly detects abnormal conditions (such as cracks, termite nests, rat holes, etc.);

[0167] S8, the comprehensive assessment module combines the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module and the image analysis module to comprehensively assess abnormal conditions;

[0168] S9, the risk prediction module integrates the monitoring results of the environmental data acquisition module, the groundwater level monitoring module, the meteorological data acquisition module, the building structure monitoring module and the image analysis module to perform multi-source data fusion and predict potential problems and risks;

[0169] S10. Alarm module: This includes an alarm that issues a warning when any abnormality or potential risk is detected. Users can view real-time monitoring data such as building settlement, tilt, and deformation through the interface.

[0170] The operating principle of a civil engineering building monitoring system disclosed herein is as follows: The image analysis module preprocesses, extracts features, and analyzes collected images to promptly identify abnormalities such as cracks, spalling, and deformation on the building surface. The infrared data analysis module analyzes infrared data to quickly detect hidden dangers such as cracks, termite nests, and rat holes within the building structure. These intelligent analysis methods, based on advanced algorithms and technologies, significantly shorten the time it takes to detect abnormalities and improve the accuracy of anomaly detection, facilitating timely action to address potential issues and ensure building safety. The comprehensive assessment module combines the results of multiple monitoring modules to comprehensively assess any abnormalities in the building. This comprehensive assessment considers the impact of multiple factors on the building, avoiding the one-sidedness of single-data assessments and enabling a more accurate assessment of the nature, extent, and impact of abnormalities, providing a reliable basis for developing appropriate treatment plans. The risk prediction module integrates multi-source data to predict potential problems and risks through comprehensive analysis of a large amount of diverse monitoring data. Multi-source data fusion can uncover correlations and underlying patterns between different data, improving the accuracy and reliability of risk prediction. Predicting potential risks in advance enables managers to take preventive measures before they occur, avoiding or minimizing losses.

[0171] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those 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 thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A civil engineering construction monitoring method, characterized in that: The following steps are involved: S1, data collection module collects civil engineering construction data, geological data and maintenance data; S2, the environmental data acquisition module collects the environmental data around the building in real time; the groundwater level monitoring module obtains the groundwater level data in real time; and collects the meteorological data of the building area in real time; S3, building structure monitoring module collects building structure data in real time; S4, the route planning module plans the optimal flight route for the drone based on the building structure and monitoring requirements; S5, the inspection module uses drones equipped with high-definition cameras and infrared scanners to conduct comprehensive monitoring of buildings along the planned route; S6, the image analysis module pre-processes, extracts features, and analyzes and identifies the collected images to identify abnormal situations; S7, the infrared data analysis module analyzes the infrared data and finds abnormal conditions; S8, the comprehensive assessment module combines data from multiple sources to comprehensively assess abnormal situations; S9, the risk prediction module integrates multi-source data and combines it with historical data to predict potential problems and risks; S10. When an abnormal situation or potential risk is detected, the alarm module will issue an alarm in time.

2. The civil engineering construction monitoring method according to claim 1, characterized in that: Step S7 includes the following steps: S71, data preprocessing: preprocessing the collected infrared data, including format conversion, denoising and temperature calibration; S72, feature extraction: including temperature feature extraction and shape and texture feature extraction; S73, abnormality identification: by setting the building material temperature threshold, monitoring the regional temperature, and marking the area as abnormal if it exceeds the threshold; S74. Build a model: Select a support vector machine (SVM) model, collect and annotate a large amount of infrared image data containing both normal and abnormal conditions, and train and test the model. Input the feature-extracted infrared image data into the trained model, and the model classifies the image based on the learned feature patterns to determine whether there is an abnormality and the type of abnormality. S75. Result verification: For the identified abnormal results, cross-validation is used to verify their accuracy; S76. Result output: Organize the infrared data analysis results into a detailed report and perform visual display.

3. The civil engineering construction monitoring method according to claim 1, characterized in that: Step S8 includes the following steps: S81, data collection: collect data from each module; S82, Data integration: Summarize data from different modules and associate them with building parts in chronological order; S83, data preprocessing: data cleaning and standardization of the integrated data; S84. Establish an evaluation indicator system: Based on professional knowledge and monitoring needs, determine key indicators and assign indicator weights; S85. Comprehensive Assessment and Analysis: Conduct multi-factor correlation analysis, applying correlation analysis to pre-processed and standardized data according to the assessment indicator system; conduct a comprehensive assessment of the integrated data based on the established assessment indicator system and weight distribution; and determine the type and degree of anomalies. S86. Generate an assessment report: Organize the results of the comprehensive assessment into a detailed report and present it visually.

4. The civil engineering construction monitoring method according to claim 3, characterized in that: In step S85, the comprehensive evaluation model is: E=[Σ n i＝1 (w i (t)x i )]+[Σ(β ij w i (t)w j (t)x i x j )]; 1≤i≤j≤n; w i (t)=w i (0)*[1+γ i *△x i (t) / (△ — x i )]*a i (t); Where, E is the comprehensive evaluation value; w i (t) is the weight of the i-th evaluation indicator at time t; x i is the value of the i-th evaluation index after standardization; x j is the jth evaluation index value after standardization; β ij is the coefficient reflecting the interaction strength between the i-th indicator and the j-th indicator; w i (t) and w j (t) are the weights of the i-th and j-th evaluation indicators at time t; x i and x j is the corresponding standardized evaluation index value; w i (0) is the initial weight; γ i is the adjustment coefficient related to the i-th indicator; △x i (t) is the difference between the value of the ith indicator at time t and the mean value of the indicator in the recent period; — x i is △x over a period of time i The average fluctuation value of a i (t) is a weight correction function based on the characteristics of the building at different stages of use; δ 1i is an adjustment parameter related to index i, which is used when the building transitions from the new construction period to the stable period of use; 2i It is an adjustment parameter related to index i, which is used when the building transitions from the stable period to the aging period; t represents the time variable; T is the design service life of the building; T1 is the dividing time point between the building's new construction period and the stable use period; T2 is the dividing time point between the building's stable use period and the aging period.

5. The civil engineering construction monitoring method according to claim 1, characterized in that: Step S9 includes the following steps: S91. Data collection and organization: Collect multi-source data and organize them according to time sequence and data type; S92. Data preprocessing: Preprocess the collected data; S93. Time series analysis; S94. Risk prediction: Combining the results of univariate and multivariate time series analysis, a risk score is assigned to each time period and indicator; S95. Finite element model construction: Select the appropriate model type based on the building's structural characteristics and potential risks. Use ANSYS professional software to build a three-dimensional model based on design drawings and actual data to simulate the interaction between the building and the actual environment. S96, finite element analysis; S97. Generate risk report: Organize the results of time series analysis and finite element analysis and prepare a detailed risk report.

6. The civil engineering construction monitoring method according to claim 5, characterized in that: Step S93 includes the following steps: S93.

1. Trend analysis: For the time series data of each monitoring indicator, the moving average method is used to analyze the long-term trend of the data; S93.

2. Seasonal analysis: Use the STL seasonal decomposition method to analyze indicators with seasonal changes and determine the influence of seasonal factors on the indicators; S93.

3. Periodicity analysis: Use Fourier transform method to identify periodic components in time series and determine the period of indicator change; S93.

4. Univariate Time Series Analysis: Build an ARIMA time series model, predict normal indicator values, and compare them with actual observed values. When the deviation between the actual value and the predicted value exceeds a threshold, mark it as a risk period. S93.

5. Multivariate time series analysis: Use the Granger causality test method to analyze the causal relationship between various indicators, screen the indicator combinations with correlation and abnormal changes, and determine the risk indicator set.

7. The civil engineering construction monitoring method according to claim 5, characterized in that: In step S94, the risk prediction model is: R(t,i)=w1(t,i)S tr (t,i)+w2(t,i)f 异 (t,i)+w3(t,i)S 因 (t,i)+S 交互 ; S 交互 =Σ[λ lm (t,i)S 因 (t,il)S 因 (t,i m )];1≤l <m≤m; S 因 (t,i)=∑ m l＝1 F ijl ;w1(t,i)+w2(t,i)+w3(t,i)=1; w1(t,i)=a1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+a2(t,i)f 异 (t,i)+a3(t,i)S 因 (t,i)]; W2(t,i)=B1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+B2(t,i)f 异 (t,i)+B3(t,i)S 因 (t,i)]; B1(t,i)+B2(t,i)+B3(t,i)=1; W3(t,i)=C1(t,i)S tr (t,i) / [a1(t,i)S tr (t,i)+C2(t,i)f 异 (t,i)+C3(t,i)S 因 (t,i)]; C1(t,i)+C2(t,i)+C3(t,i)=1; where R(t,i) is the value of the risk score function; t represents the time period; i refers to various monitoring indicators obtained from multi-source data collection; w1(t,i) reflects the trend change degree score S tr The relative importance of (t,i) in the risk score; w2(t,i) reflects the frequency of outliers f 异 (t,i); w3(t,i) represents the causal relationship strength score S 因 The weight of (t,i); S tr (t,i) is the trend change degree score; f 异 (t,i) is the frequency of occurrence of outliers; S 因 (t,i) is the causal relationship strength score; F ijl is the F value of the Granger causality test between index i and jl; lm (t,i) is the coefficient reflecting the strength of the interaction between the causal relationship between indicator il and indicator im; α1(t,i) is the coefficient used to adjust the weight of the trend change degree score in w1(t,i); a2(t,i) is the coefficient used to adjust the relative weight of the frequency of abnormal values ​​in the calculation of w1(t,i); a3(t,i) is the coefficient used to adjust the relative weight of the causal relationship strength score between indicators; B1(t,i) is the coefficient used to adjust the relative weight of the trend change degree score in the calculation of w2(t,i); B2(t,i) is the coefficient used to adjust the weight of the frequency of occurrence of outliers in w2(t,i); B3(t,i) is the coefficient used to adjust the relative weight of the causal relationship strength score between indicators; C1(t,i) is the coefficient used to adjust the relative weight of the trend change degree score in the calculation of w3(t,i); C2(t,i) is the coefficient used to adjust the relative weight of the frequency of occurrence of outliers in the calculation of w3(t,i); C3(t,i) is the coefficient used to adjust the weight of the causal relationship strength score in w3(t,i).

8. The civil engineering construction monitoring method according to claim 5, characterized in that: Step S96 includes the following steps: S96.

1. Load application: Determine the load conditions for finite element analysis based on the potential risk periods and indicators determined by time series analysis and in combination with the actual project conditions; S96.

2. Simulation Analysis: Run finite element analysis software to simulate the mechanical response of the building structure under different load conditions and analyze the mechanical behavior of the structure under the influence of potential risk factors; S96.

3. Result evaluation: Based on the finite element simulation results, evaluate the specific impact of potential risks on the building structure.

9. The civil engineering construction monitoring method according to claim 1, characterized in that: Step S6 includes the following steps: S61. Image acquisition: acquiring relevant images of civil engineering buildings; annotating the acquired images; S62, image preprocessing: preprocessing the collected image; S63, feature extraction: extracting features from the preprocessed image, where the extracted features include color, texture, and shape; S64. Build a model: Build a convolutional neural network model, train the model using the training set, evaluate and adjust parameters using the validation set, evaluate model performance using the test set, and continuously optimize the model based on the evaluation results. S65, analysis and recognition: The pre-processed and feature-extracted image is input into the trained model, and the model classifies the image to determine whether there is an anomaly and the type of anomaly; S66. Result output: output the identified abnormal situation in an intuitive manner.

10. A civil engineering construction monitoring system comprising: Control center, comprehensive assessment module: integrates environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, risk prediction module, alarm module, building structure monitoring module and image analysis module; its characteristics are: Data collection module: collects civil engineering construction data, geological data and maintenance data; annotates the data as a reference sample; Route planning module: plans the optimal flight route for drones based on the building’s structure and monitoring requirements; Inspection module: including drones equipped with high-definition cameras and infrared scanners to conduct all-round monitoring of buildings; Environmental data acquisition module: real-time collection of building surrounding environment data; Groundwater level monitoring module: obtain groundwater level data in real time; Meteorological data collection module: collects meteorological data of the construction area in real time; Building structure monitoring module: deploy multiple sensors at multiple locations on the building to collect building structure data in real time; Image analysis module: pre-processes, extracts features, and analyzes and identifies collected images to promptly identify abnormalities; Infrared data analysis module: analyze infrared data and detect abnormal situations in time; Comprehensive assessment module: combines the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to comprehensively assess abnormal situations; Risk prediction module: Integrates the monitoring results of the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, building structure monitoring module and image analysis module to perform multi-source data fusion and predict potential problems and risks; Alarm module: including alarm, which will issue an alarm in time when abnormal conditions or potential risks are detected; Interactive module: Develop a user interactive interface with complete functions and convenient operation, through which users can view monitoring data in real time; Control center: and comprehensive assessment module: connect the environmental data acquisition module, groundwater level monitoring module, meteorological data acquisition module, risk prediction module, alarm module, building structure monitoring module and image analysis module to the network.

Images