An automated measurement system and method for deformation monitoring in civil engineering projects.

Abstract

This application provides an automated measurement system and method for monitoring deformation in civil engineering projects. It determines visual and physical feature vector sequences using time-series visual and physical quantity data of the monitoring area within a target civil engineering structure. It then determines the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure, fusing the visual and physical feature vector sequences into a fused deformation feature sequence characterizing the monitoring area. Based on the structural design parameter set of the target civil engineering structure and the fused deformation feature sequence, it determines the predicted deformation trend of the monitoring area. Finally, it provides early warning of the deformation state of the monitoring area based on the predicted deformation trend. Using the scheme of this application, deformation in civil engineering projects can be monitored based on the correlation between structural deformation and visual and physical quantity characteristics.

Description

Technical Field

[0001] This application relates to the field of automated measurement technology, and more specifically, to an automated measurement system and method for monitoring deformation in civil engineering projects. Background Technology

[0002] Automated measurement refers to the technical process of automatically completing data acquisition, analysis, and feedback by integrating sensors, controllers, and data processing systems. It replaces traditional manual operation, achieves high-precision and high-efficiency continuous monitoring, and is widely used in industrial manufacturing, scientific research, environmental monitoring, and other fields. It significantly improves the reliability, consistency, and real-time performance of measurements and is one of the core supporting technologies for intelligent manufacturing and digital management.

[0003] In the current field of civil engineering structural deformation monitoring, traditional monitoring methods mostly rely on manual on-site measurement or data collection using single-type sensors, resulting in low monitoring efficiency, poor data timeliness, and high labor costs. While some automated monitoring solutions can acquire visual and physical data separately, they generally analyze these two types of data independently, failing to establish a quantitative correlation between structural deformation and visual and physical characteristics. This leads to a lack of core theoretical support for multi-source data fusion, and the fusion results cannot fully reflect the true deformation state of the structure. At the same time, existing deformation trend prediction methods mostly rely on simple time-series fitting based on monitoring data, without considering the inherent properties of the structural design parameter set for correction. Consequently, the prediction results deviate significantly from the actual service state of the structure. Therefore, how to monitor civil engineering deformation based on the correlation between structural deformation and visual and physical characteristics has become a problem facing the industry. Summary of the Invention

[0004] This application provides an automated measurement system and method for monitoring deformation in civil engineering projects, which can monitor the deformation of civil engineering projects based on the correlation between structural deformation and visual and physical characteristics.

[0005] In a first aspect, this application provides an automated measurement method for deformation monitoring in civil engineering, comprising the following steps:

[0006] Collect time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure;

[0007] A sequence of visual feature vectors characterizing the structural deformation of the monitoring area is extracted from the time-series visual data, and a sequence of physical feature vectors is extracted from the time-series physical quantity data;

[0008] Based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined. The visual feature vector sequence and the physical feature vector sequence are cross-fused through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area.

[0009] Obtain the set of structural design parameters for the target civil engineering structure, and predict and analyze the deformation trend of the monitored area based on the set of structural design parameters and the fused deformation feature sequence to obtain the predicted deformation trend of the monitored area.

[0010] Based on the predicted deformation trend, an early warning is issued for the deformation status of the monitored area in the target civil engineering structure.

[0011] In some embodiments, extracting a sequence of visual feature vectors characterizing the structural deformation of the monitored area from the temporal visual data specifically includes:

[0012] The temporal visual data is preprocessed to obtain preprocessed temporal visual data;

[0013] Deformation features are extracted from the preprocessed temporal visual data to obtain a sequence of visual feature vectors characterizing the structural deformation of the monitored area.

[0014] In some embodiments, extracting a sequence of physical feature vectors from the time-series physical quantity data specifically includes:

[0015] The time-series physical quantity data is preprocessed to obtain preprocessed time-series physical quantity data;

[0016] Physical quantity features are extracted from the preprocessed time-series physical quantity data to obtain a sequence of physical feature vectors.

[0017] In some embodiments, determining the coupling relationship between deformation and visual and physical quantities in a target civil engineering structure based on historical monitoring data of deformation-related visual and physical quantities specifically includes:

[0018] Acquire historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure;

[0019] Determine the correlation coefficients between the parameters in the historical monitoring data;

[0020] The coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined based on all correlation coefficients and the historical monitoring data.

[0021] In some embodiments, the cross-fusion of the visual feature vector sequence and the physical feature vector sequence through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area specifically includes:

[0022] The visual feature vector sequence and the physical feature vector sequence are time-aligned to obtain the aligned visual feature vector sequence and physical feature vector sequence;

[0023] Based on the aforementioned coupling relationship, the aligned visual feature vector sequence and physical feature vector sequence are weighted and fused according to their corresponding dimensions to obtain fused feature vectors for each dimension;

[0024] All the fused feature vectors are cross-stacked to obtain a fused deformation feature sequence characterizing the monitored area.

[0025] In some embodiments, predicting and analyzing the deformation trend of the monitored area based on the structural design parameter set and the fused deformation feature sequence to obtain the predicted deformation trend of the monitored area specifically includes:

[0026] The structural design parameter set and the fused deformation feature sequence are standardized to obtain a standardized structural design parameter set and fused deformation feature sequence.

[0027] The standardized fused deformation feature sequence is fitted with deformation trends at multiple time scales to obtain the preliminary deformation trend of the monitored area.

[0028] The preliminary deformation trend is corrected based on the standardized set of structural design parameters to obtain the predicted deformation trend of the monitored area.

[0029] In some embodiments, providing early warning of the deformation state of a monitored area in a target civil engineering structure based on the predicted deformation trend specifically includes:

[0030] Extract various early warning indicators from the predicted deformation trend;

[0031] Establish a three-level early warning threshold system for target civil engineering structures;

[0032] Each early warning indicator is compared with the three-level early warning threshold system to provide early warning of the deformation status of the monitored area in the target civil engineering structure.

[0033] Secondly, this application provides an automated measurement system for deformation monitoring in civil engineering projects, comprising:

[0034] The acquisition module is used to acquire time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure;

[0035] The processing module is used to extract a sequence of visual feature vectors characterizing the structural deformation of the monitoring area from the time-series visual data, and to extract a sequence of physical feature vectors from the time-series physical quantity data.

[0036] The processing module is also used to determine the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, and to cross-fuse the visual feature vector sequence and the physical feature vector sequence through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area.

[0037] The processing module is also used to obtain the structural design parameter set of the target civil engineering structure, and to predict and analyze the deformation trend of the monitoring area based on the structural design parameter set and the fused deformation feature sequence, so as to obtain the predicted deformation trend of the monitoring area.

[0038] The execution module is used to provide early warning of the deformation status of the monitored area in the target civil engineering structure based on the predicted deformation trend.

[0039] Thirdly, this application provides a computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described automated measurement method for deformation monitoring in civil engineering.

[0040] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described automated measurement method for deformation monitoring in civil engineering.

[0041] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0042] The automated measurement system and method for deformation monitoring in civil engineering provided in this application avoids the limitation of a single data dimension failing to fully reflect the correlation between deformation and features by collecting time-series visual and physical quantity data. It extracts two types of feature vector sequences, transforming the original data into a core carrier that accurately represents deformation, giving the correlation analysis between deformation and features a clear target and solving the problem of disorganized original data making it difficult to establish effective correlations. By determining the coupling relationship through historical data and achieving cross-fusion, it directly establishes a quantitative mapping logic between deformation and the two features, allowing the fusion result to fully integrate the correlation information between the two types of features and deformation, overcoming the defects of traditional methods where independent analysis of two features leads to broken correlations and one-sided deformation representations. Combining the structural design parameter set with the predicted trend of the fused feature sequence, it utilizes the inherent properties of the structure to optimize the application accuracy of the correlation relationship, ensuring that the deformation trend prediction always anchors to the core correlation between deformation and features, improving the matching degree of the prediction results to the actual deformation state. Based on the predicted trend warning, it transforms the correlation law between deformation and features into a basis for safety judgment, realizing a closed loop from correlation analysis to risk warning, ensuring that the warning result accurately correlates with the essential characteristics of deformation, and solving the problem of judgment distortion caused by traditional warnings being detached from feature correlations. Using the above scheme, the deformation of civil engineering structures can be monitored based on the correlation between structural deformation and visual and physical characteristics. Attached Figure Description

[0043] Figure 1 This is an exemplary flowchart of an automated measurement method for deformation monitoring in civil engineering, according to some embodiments of this application;

[0044] Figure 2 This is an exemplary flowchart illustrating the determination of coupling relationships according to some embodiments of this application;

[0045] Figure 3 This is an exemplary flowchart illustrating the determination of predicted deformation trends according to some embodiments of this application;

[0046] Figure 4 This is a structural schematic diagram of an automated measurement system for deformation monitoring in civil engineering, according to some embodiments of this application;

[0047] Figure 5 This is a structural schematic diagram of a computer device for implementing an automated measurement method for deformation monitoring in civil engineering, according to some embodiments of this application. Detailed Implementation

[0048] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] refer to Figure 1The figure is an exemplary flowchart of an automated measurement method for civil engineering deformation monitoring according to some embodiments of this application. The automated measurement method for civil engineering deformation monitoring mainly includes the following steps:

[0050] In step 101, time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure are collected.

[0051] It should be noted that the temporal visual data in this application refers to a sequence of visual data continuously collected from the monitoring area of ​​the civil engineering structure at a preset time sampling interval. The temporal visual data includes a high-definition image sequence, a three-dimensional point cloud sequence, and a structural surface texture image sequence of the monitoring area, reflecting the dynamic changes over time in the geometric contour morphology, spatial location of feature points, development state of surface cracks, and intuitive representation of relative displacement of components in the monitoring area. The temporal physical quantity data refers to a sequence of physical parameter data continuously collected from the monitoring area and its surrounding environment at a set time frequency. The temporal physical quantity data includes data on structural strain, displacement, stress, vibration frequency, ambient temperature and humidity, and wind pressure within the monitoring area, reflecting the evolution over time of the mechanical response state of the structure, the effect of load, the degree of influence of environmental factors, and the physical characteristics closely related to structural deformation in the monitoring area.

[0052] In step 102, a sequence of visual feature vectors characterizing the structural deformation of the monitoring area is extracted from the time-series visual data, and a sequence of physical feature vectors is extracted from the time-series physical quantity data.

[0053] In some embodiments, extracting a sequence of visual feature vectors characterizing the structural deformation of the monitored area from the temporal visual data can be achieved using the following steps:

[0054] The temporal visual data is preprocessed to obtain preprocessed temporal visual data;

[0055] Deformation features are extracted from the preprocessed temporal visual data to obtain a sequence of visual feature vectors characterizing the structural deformation of the monitored area.

[0056] In specific implementation, the temporal visual data is preprocessed to obtain the preprocessed temporal visual data. This can be achieved in the following ways: For the image sequence in the temporal visual data, a 3×3 or 5×5 Gaussian filter kernel is used to perform convolution operations on the image to filter out random noise. The grayscale image of the monitoring area is stretched using the histogram equalization method to correct the problem of uneven grayscale distribution. Then, the segmentation threshold is determined based on the maximum inter-class variance method, and the threshold segmentation method is used to accurately segment the target monitoring area from the complex background. For the three-dimensional point cloud sequence in the temporal visual data, a statistical filtering algorithm is used to calculate the mean and variance of the distances within the neighborhood of each point. Outliers exceeding three times the variance are removed. The point cloud data at the initial time is selected as the reference point cloud. The iterative nearest point algorithm is used to iteratively calculate the optimal rigid body transformation matrix of the corresponding point pairs to complete the coordinate registration of the point cloud data at different time nodes, so that the point cloud data at each time node are in the same world coordinate system. This is how the preprocessed temporal visual data is obtained. Other methods can also be used in other embodiments, which are not limited here.

[0057] In addition, in specific implementation, the deformation feature extraction of the preprocessed temporal visual data to obtain the visual feature vector sequence representing the structural deformation of the monitored area can be achieved in the following way: For the preprocessed image sequence in the preprocessed temporal visual data, a Gaussian difference pyramid is constructed using the scale-invariant feature transform algorithm and extreme points are detected. After determining the main direction of the key points, feature descriptors are generated. Corresponding key points at adjacent time nodes are matched. The pixel coordinate deviation value is converted into the actual physical displacement value by combining the camera calibration parameters. At the same time, the Cannibal edge detection operator is used to sequentially perform Gaussian smoothing, gradient magnitude and direction calculation, non-maximum suppression, double threshold detection and edge connection operations to extract the structural contour of the monitored area. The contour matching algorithm is used to calculate the Hausdorff distance and angular offset of the contours at adjacent time points. For the preprocessed three-dimensional image sequence in the preprocessed temporal visual data, the deformation feature extraction is performed in the following way: The n-dimensional point cloud sequence is obtained by selecting 15 neighborhood points for each sampling point, calculating the neighborhood curvature and normal vector by calculating the covariance matrix of the neighborhood point set, and statistically analyzing the rate of change of curvature and the angle between the normal vectors of sampling points at adjacent time points. At the same time, the centroid coordinate deviation value and the point cloud volume change rate are calculated. Then, the z-score normalization method is used to map the above-mentioned feature point displacement, contour offset parameter, point cloud curvature change rate, normal vector angle, centroid deviation value, volume change rate and other parameters related to structural deformation to the interval [-1,1]. The multi-dimensional feature parameters corresponding to each time node are integrated into an n-dimensional visual feature vector according to the preset dimension order. Then, the feature vectors of all time nodes are sorted according to the time sampling order to finally obtain the visual feature vector sequence representing the structural deformation of the monitored area. Other methods can be used in other embodiments, which are not limited here.

[0058] It should be noted that the visual feature vector sequence in this application represents a sequence of temporally ordered multidimensional visual features of the target civil engineering structure deformation monitoring area. The visual feature vector sequence reflects the dynamic evolution process related to deformation, such as the spatial position offset of structural feature points, changes in structural outline morphology, surface crack initiation and propagation, and three-dimensional geometric distortion within the monitoring area. It can be used to analyze the deformation state of the target civil engineering structure.

[0059] In some embodiments, the extraction of physical feature vector sequences from the time-series physical quantity data can be achieved by the following steps:

[0060] The time-series physical quantity data is preprocessed to obtain preprocessed time-series physical quantity data;

[0061] Physical quantity features are extracted from the preprocessed time-series physical quantity data to obtain a sequence of physical feature vectors.

[0062] In specific implementation, the time-series physical quantity data is preprocessed to obtain the preprocessed time-series physical quantity data. This can be achieved in the following way: For different types of physical quantity data such as strain, displacement, stress, vibration frequency, and ambient temperature and humidity in the time-series physical quantity data, a 5-point moving average filtering algorithm is used to perform convolution operation on the data sequence to filter out random errors introduced by sensor inherent noise and environmental electromagnetic interference. The significance level is set to 0.05 using the Grubbs criterion. The residuals of each data point are calculated and compared with the critical value to eliminate abnormal data points caused by sensor instantaneous failure or load change. At the same time, data range calibration and timestamp synchronization calibration are completed according to the calibration curves of each sensor to ensure the consistency and numerical accuracy of different types of physical quantity data in the time dimension. This is how the preprocessed time-series physical quantity data is obtained. Other methods can also be used in other embodiments, which are not limited here.

[0063] In addition, in specific implementation, the physical feature extraction of the preprocessed time-series physical quantity data to obtain the physical feature vector sequence can be achieved in the following way: First, based on the principles of civil engineering structural mechanics, select the physical parameters in the preprocessed time-series physical quantity data that are strongly correlated with structural deformation, and remove redundant parameters with low correlation to deformation. Then, calculate the derived features of each parameter in the preprocessed time-series physical quantity data. These derived features specifically include strain change gradient, cumulative displacement increment, stress fluctuation amplitude, vibration frequency offset, and the mean, variance, peak value, and trough value of each parameter within a 10-minute time window. Statistical characteristics such as values ​​are analyzed, and then the min-max normalization method is used to map the values ​​of each parameter and derived feature in the time series physical quantity data to the [0,1] interval, eliminating the interference of differences in physical dimensions on subsequent analysis. Finally, according to the order of time sampling, the multi-dimensional physical feature parameters corresponding to each time node are integrated into an n-dimensional feature vector according to the preset dimension arrangement rules. The feature vectors of all time nodes are sorted in sequence to form a physical feature vector sequence that can quantitatively reflect the dynamic evolution law of the physical characteristics related to structural deformation in the monitoring area. Other methods can be used in other embodiments, which are not limited here.

[0064] It should be noted that the physical feature vector sequence in this application represents the sequence of temporal multidimensional physical features of the deformation monitoring area of ​​the target civil engineering structure. It reflects the mechanical response state, displacement evolution law, changes in natural vibration characteristics, and the degree of influence of environmental factors on the physical properties of the structure within the monitoring area. It can be used to analyze the deformation state of the target civil engineering structure.

[0065] In step 103, based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined. The visual feature vector sequence and the physical feature vector sequence are cross-fused through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area.

[0066] In some embodiments, reference Figure 2 The figure is an exemplary flowchart for determining coupling relationships in some embodiments of this application. In this embodiment, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure can be determined based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure using the following steps:

[0067] In step 1031, historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure are obtained;

[0068] In step 1032, the correlation coefficients between the parameters in the historical monitoring data are determined;

[0069] In step 1033, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined based on all correlation coefficients and the historical monitoring data.

[0070] In practice, obtaining historical monitoring data of deformation-related visual and physical quantities in the target civil engineering structure can be achieved as follows: Retrieve a complete historical dataset of the target civil engineering structure under the same monitoring area and working conditions. This dataset specifically includes a historical sequence of time-series visual features, a historical sequence of time-series physical quantity features, and the measured structural deformation data within the corresponding time period. Then, preprocess the three types of data. First, accurately align the timeline based on the monitoring timestamp to ensure a one-to-one correspondence between visual features, physical quantity parameters, and deformation data at the same moment. For the few missing points in the data sequence, use linear interpolation. The value method completes the data based on the numerical change trend of the two adjacent valid data points before and after the missing point. Then, the significance level is set to 0.05 using the Grubbs criterion. First, the overall dispersion of the entire set of data is calculated. Then, each data point is judged to determine whether the deviation from the overall data distribution exceeds the critical range. Abnormal data points that exceed the critical range are removed. Finally, all visual feature parameters and physical quantity parameters are standardized to map parameters of different dimensions to the same numerical range, ensuring data format uniformity and reliability, and obtaining preprocessed historical monitoring data. Other methods can be used in other embodiments, which are not limited here.

[0071] In addition, in specific implementation, the correlation coefficient between various parameters in the historical monitoring data can be determined in the following way: Select the corresponding analysis method for the correlation characteristics between different types of characteristic parameters and deformation in the historical monitoring data. For linear correlation characteristic parameters whose data distribution conforms to a normal distribution, such as the correlation between strain, displacement and deformation, the Pearson correlation coefficient analysis method is used to calculate the correlation coefficient. First, the overall central tendency and dispersion of each characteristic parameter and the measured deformation data are statistically analyzed. Then, the linear correlation strength is quantified by analyzing the synchronicity of the changes in the two data, i.e., the correlation coefficient. For nonlinear correlation characteristic parameters whose data distribution does not conform to a normal distribution, such as the correlation between environmental temperature and humidity and deformation, the Spearman rank correlation coefficient analysis method is used to calculate the correlation coefficient. First, the two types of data are ranked according to their numerical values. Then, the nonlinear correlation strength is measured by calculating the distribution law of the rank difference after ranking, i.e., the correlation coefficient. Other methods can also be used in other embodiments, which are not limited here.

[0072] Furthermore, in specific implementation, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure can be determined based on all correlation coefficients and the historical monitoring data in the following manner: All correlation coefficients are organized into a two-dimensional correlation matrix, clearly presenting the degree of correlation between each feature parameter and deformation, and the feature parameters themselves. A correlation coefficient absolute value of 0.7 is set as the high correlation screening threshold. Visual feature parameters and physical quantity parameters with values ​​greater than this threshold in the correlation matrix are retained as high-correlation feature parameters from the historical monitoring data, while low-correlation redundant parameters with values ​​less than this threshold are removed to reduce the computational complexity of subsequent modeling. The selected high-correlation visual feature parameters and physical quantity parameters are used as input variables, and the corresponding measured structural deformation data are used as output variables. A multiple linear regression model is used to construct the basic framework of the coupling relationship. The preprocessed historical monitoring data is divided into a training set and a test set in a 7:3 ratio, where the training set is used to solve the model parameters, and the test set is used to verify the model accuracy. The least squares method is then used to process the training set data. Iterative calculations are performed, continuously adjusting the regression coefficients and constant terms in the model to minimize the deviation between the predicted deformation output and the measured deformation in the training set. The coefficient of determination and root mean square error (RMSE) are used as evaluation metrics for model accuracy. The coefficient of determination measures the model's ability to explain data variation patterns, while the RMSE measures the average deviation between predicted and measured values. Test set data is input into the model. If the coefficient of determination is less than 0.85, the model's fit is not meeting the preset requirements. In this case, a polynomial term is introduced to optimize the model, enhancing its ability to fit nonlinear relationships. The optimized model parameters are then iteratively solved using the training set data again, and the accuracy is verified again using the test set. This iterative process is repeated until the model meets the preset requirements, ultimately yielding a coupled relationship model that quantifies the mapping relationship between structural deformation and visual features and physical quantities. This coupled relationship model is used as the coupled relationship between deformation and visual and physical quantities in the target civil engineering structure. Other implementation methods can also be used in other embodiments, which are not limited here.

[0073] It should be noted that the historical monitoring data in this application represents the historical basic data source of the target civil engineering structure deformation monitoring area, reflecting the historical evolution of the structure's visual characteristics, physical properties, and deformation state under different working conditions and time dimensions; the correlation coefficient represents the coefficient of the correlation strength of the civil engineering structure deformation monitoring area, reflecting the close correlation between various characteristic parameters and structural deformation; the coupling correlation relationship represents the characteristic and deformation mapping law of the civil engineering structure deformation monitoring area, reflecting the static correspondence and dynamic mapping law between multi-source characteristic parameters and structural deformation.

[0074] In some embodiments, the cross-fusion of the visual feature vector sequence and the physical feature vector sequence through the coupling correlation to obtain a fused deformation feature sequence characterizing the monitoring area can be achieved by the following steps:

[0075] The visual feature vector sequence and the physical feature vector sequence are time-aligned to obtain the aligned visual feature vector sequence and physical feature vector sequence;

[0076] Based on the aforementioned coupling relationship, the aligned visual feature vector sequence and physical feature vector sequence are weighted and fused according to their corresponding dimensions to obtain fused feature vectors for each dimension;

[0077] All the fused feature vectors are cross-stacked to obtain a fused deformation feature sequence characterizing the monitored area.

[0078] In specific implementation, the visual feature vector sequence and the physical feature vector sequence are time-aligned to obtain the aligned visual feature vector sequence and physical feature vector sequence. This can be achieved as follows: using the monitoring timestamps of the visual feature vector sequence and the physical feature vector sequence as a unified benchmark, all valid time nodes in the visual feature vector sequence and the physical feature vector sequence are first extracted and a time index table is established. The time indices of the two types of sequences are compared. If there is a missing time node on one side, linear interpolation is used to accurately fill it in. Specifically, based on the feature values ​​of the two valid nodes adjacent to the missing node, the feature change rate between the two points is calculated. Combined with the time interval between the missing node and the previous valid node, the feature value of the missing node is deduced. The feature values ​​are calculated as follows: If duplicate timestamp nodes exist, the latest collected feature vector data is retained to avoid redundancy; at the same time, the Grubbs criterion is used to remove outlier data, with a significance level of 0.05. First, the mean and dispersion of each vector in the two types of feature vector sequences are calculated, and then the deviation of each feature vector from the mean is calculated one by one. Outlier vectors with deviations exceeding the critical range are removed. Then, the remaining feature vectors are subjected to secondary timestamp calibration to ensure that there are one-to-one corresponding visual feature vectors and physical feature vectors under the same time index. Finally, the visual feature vector sequence and physical feature vector sequence with complete time dimension matching and reliable data alignment are obtained. Other methods can be used in other embodiments, which are not limited here.

[0079] Furthermore, in specific implementation, based on the aforementioned coupling relationship, the aligned visual feature vector sequence and physical feature vector sequence are weighted and fused according to their corresponding dimensions to obtain the fused feature vectors for each dimension. This can be achieved in the following way: Based on a pre-determined coupling relationship model, the association weights between each visual feature dimension, physical feature dimension, and structural deformation amount, which have been calibrated using historical data, are extracted from the model. These weights are the regression coefficients corresponding to each feature parameter within the model. The larger the absolute value of the weight, the higher the contribution of that feature dimension to the representation of the structural deformation state. To optimize computational efficiency, a weight selection threshold of 0.6 is set, and two types of sequences are retained through feature dimension index matching. For feature dimensions whose absolute weight is greater than the threshold, low-weight redundant dimensions are removed, and the retained feature dimensions are numbered and marked to ensure that the correspondence between visual and physical feature dimensions is clear and traceable. Then, for the visual feature vector and physical feature vector of each time node in the aligned sequence, a weighted operation is performed one by one according to the marked feature dimension. That is, the value of each visual feature dimension is multiplied by its corresponding association weight, and the value of each physical feature dimension is also multiplied by its corresponding association weight to obtain the weighted visual feature value and weighted physical feature value of each feature dimension, so as to obtain the fused feature vector of each dimension. Other methods can be used in other embodiments, which are not limited here.

[0080] Furthermore, in specific implementation, the cross-superposition of all fused feature vectors to obtain the fused deformation feature sequence representing the monitoring area can be achieved in the following way: The weighted visual feature values ​​and weighted physical feature values ​​corresponding to the number at the same time node are numerically superimposed. For example, the weighted value of the "feature point displacement dimension" in the visual features is superimposed with the weighted value of the "displacement sensor monitoring dimension" in the physical features. All superimposed dimension values ​​are integrated according to a preset dimension order to form a fused feature vector for a single time node. To eliminate potential dimensional differences after superimposing features of different dimensions, the z-score normalization method is used to uniformly map the values ​​of all fused feature vectors to the [-1,1] interval, ensuring that the features of each dimension are consistent. Weight balancing is performed. Then, according to the order of time sampling, the fused feature vectors of all time nodes are arranged in order from early to late according to the time index to form an initial fused feature sequence. Finally, a 3-point moving average filtering algorithm is used to smooth the initial sequence. That is, with each fused feature vector as the center, one adjacent vector before and after it is selected to form a sliding window. The arithmetic mean of the corresponding dimension values ​​of the three vectors in the window is calculated and replaced with the value of the original center vector. The entire sequence is smoothed point by point to filter out random noise introduced during the fusion process. Finally, a fused deformation feature sequence that can fully integrate visual and physical dual-dimensional deformation information and characterize the dynamic evolution state of structural deformation in the monitoring area is obtained. Other methods can be used in other embodiments, which are not limited here.

[0081] It should be noted that the fusion feature vector in this application represents a single-time-node multidimensional deformation comprehensive characterization unit of the target civil engineering structure deformation monitoring area, reflecting the comprehensive situation of the geometric morphological changes and mechanical response state of the structural deformation in the monitoring area at the corresponding time node; the fusion deformation feature sequence belongs to the time-series structural deformation comprehensive characterization set in this field, reflecting the dynamic evolution law of structural deformation in the monitoring area from the past to the present, and can be used to judge and analyze the deformation state of the target civil engineering structure.

[0082] In step 104, the structural design parameter set of the target civil engineering structure is obtained, and the deformation trend of the monitoring area is predicted and analyzed based on the structural design parameter set and the fused deformation feature sequence to obtain the predicted deformation trend of the monitoring area.

[0083] It should be noted that the structural design parameter set in this application represents the set of basic technical parameters for the design of the target civil engineering structure. It is a key parameter system supporting structural scheme design, mechanical performance verification, and deformation trend analysis. It reflects the inherent properties of the structure, the mechanical properties of materials, design basis conditions, and performance control requirements. It is an important benchmark for judging whether the actual deformation of the structure exceeds the safety threshold. The structural design parameter set includes structural geometric parameters, material performance parameters, mechanical design parameters, structural design parameters, and safety control parameters. Among them, structural geometric parameters include member length, cross-sectional dimensions, span, story height, member spacing, and overall structural geometry. Material performance parameters include concrete strength grade, steel yield strength, steel elastic modulus, material Poisson's ratio, and material shear strength. Mechanical design parameters include standard values ​​of design dead load, standard values ​​of live load, load combination coefficient, seismic fortification intensity, and wind load values. Structural design parameters include member reinforcement ratio, node connection form, steel anchorage length, embedded part specifications and placement. Safety control parameters include structural design deformation limits, warning thresholds corresponding to different safety levels, crack resistance control indicators, and structural safety reserve coefficients.

[0084] In some embodiments, reference Figure 3 The figure is an exemplary flowchart for determining the predicted deformation trend in some embodiments of this application. In this embodiment, the predicted deformation trend of the monitoring area is predicted and analyzed based on the structural design parameter set and the fused deformation feature sequence. The predicted deformation trend of the monitoring area can be obtained by the following steps:

[0085] In step 1041, the structural design parameter set and the fused deformation feature sequence are standardized to obtain a standardized structural design parameter set and fused deformation feature sequence.

[0086] In step 1042, the standardized fused deformation feature sequence is fitted with deformation trends at multiple time scales to obtain the preliminary deformation trend of the monitored area;

[0087] In step 1043, the preliminary deformation trend is corrected based on the standardized set of structural design parameters to obtain the predicted deformation trend of the monitoring area.

[0088] In specific implementation, the standardized structural design parameter set and the fused deformation feature sequence are obtained by standardizing the structural design parameter set and the fused deformation feature sequence. This can be achieved in the following way: First, perform dimensional mapping on the structural design parameter set and the fused deformation feature sequence to clarify the one-to-one matching relationship between the deformation parameters of each dimension in the fused deformation feature sequence and the corresponding reference parameters in the structural design parameter set. Specifically, displacement-type fused parameters match design deformation limits, strain-type fused parameters match material yield strain values, stress-type fused parameters match design stress standard values, and geometric shape-type fused parameters match component cross-sectional dimensions and arrangement spacing. Then, the min-max standardization method is used to perform parameter-by-parameter standardization on both types of parameters. The process involves calculating the global maximum and minimum values ​​of each parameter within the entire dataset, then mapping each parameter value to the [0,1] interval. Simultaneously, the Grubbs criterion is used to remove outliers with a significance level of 0.05. Specifically, the mean and standard deviation of each parameter sequence are calculated first, and then the absolute value of the residuals of each data point is checked against the Grubbs critical value. Outliers exceeding the critical value are removed. After removal, the parameter sequence is checked for integrity to ensure there are no consecutive missing data segments. Finally, a standardized set of structural design parameters and a fused deformation feature sequence with unified dimensions and reliable data are obtained. Other implementation methods can also be used in other embodiments, which are not limited here.

[0089] In addition, in specific implementation, the deformation trend fitting of the standardized fused deformation feature sequence at multiple time scales to obtain the preliminary deformation trend of the monitoring area can be achieved in the following way: using an autoregressive moving average model as the fitting tool, firstly, the optimal order of the model is determined by the Akaike Information Criterion. Specifically, different combinations of autoregressive order p (1-5) and moving average order q (1-5) are traversed, the Akaike Information Criterion value corresponding to each combination is calculated, and the p and q combination with the smallest Akaike Information Criterion value is selected as the optimal order of the model; then, according to the actual needs of structural deformation monitoring, the standardized fused deformation feature sequence is divided into three subsequences according to the time scale, where the hourly subsequence corresponds to short-term deformation monitoring, the daily subsequence corresponds to daily operation deformation monitoring, and the weekly subsequence corresponds to long-term deformation trend monitoring. Data segmentation is performed on each subsequence to ensure that the data of each subsequence is continuous and the time interval is uniform; then, each subsequence is substituted into the autoregressive moving average model of the optimal order for training, and the least squares method is used to iteratively solve the model parameters. The iteration convergence condition is set as the parameter change is less than 10. -5 After the model training is completed, the model is used to fit the trend of each subsequence, and the trend index of deformation parameters in each subsequence is calculated, including the deformation change rate, the cumulative deformation increment, and the peak occurrence period. The fitting results of three time scales, namely hourly, daily, and weekly, are integrated to obtain the preliminary deformation trend that can cover the monitoring area in the short, medium and long term. Other methods can be used in other embodiments, which are not limited here.

[0090] In addition, in specific implementation, the predicted deformation trend of the monitoring area can be obtained by correcting the preliminary deformation trend based on the standardized structural design parameter set in the following way: extracting each correction coefficient from the standardized structural design parameter set, including load combination coefficient, environmental adaptability coefficient, material aging correction coefficient, and safety reserve coefficient. Among them, the load combination coefficient is used to compensate for the deviation between the actual load and the design standard load; the environmental adaptability coefficient is used to correct the influence of environmental factors such as temperature, humidity, and wind pressure on structural deformation; the material aging correction coefficient is used to consider the deformation deviation caused by the degradation of material properties such as concrete carbonation and steel corrosion; and the safety reserve coefficient is used to reflect the inherent redundancy of the structure. The constraint effect of load capacity on deformation development is specifically extracted as follows: The load combination coefficient is calculated based on the standard values ​​of design dead load, live load, and load partial factors in the design parameter set, referencing the combination rules for different load conditions, and combining the actual service load types in the monitoring area. This coefficient is used to match the deviation between the actual load and the design standard load. The environmental adaptability coefficient is based on the structural environment category and material weathering resistance level in the design parameter set, comparing the influence weights of different environments on structural deformation, and quantifying it using real-time temperature, humidity, and wind pressure monitoring data of the monitoring area. The material aging correction coefficient is based on the structural design service life, concrete strength grade, and steel reinforcement cover thickness in the design parameter set, using material... An empirical model for material performance degradation is used, calculated by combining the correlation between concrete carbonation depth, steel corrosion rate, and design parameters, to consider the impact of material performance degradation on deformation. The safety reserve coefficient is determined based on the structural safety level, component bearing capacity design value, and material strength design value in the design parameter set, according to the structural reliability design principle, by calculating the ratio of design bearing capacity to ultimate bearing capacity, thus reflecting the constraint effect of the structure's inherent redundancy on deformation development. For the predicted values ​​at different time scales in the initial deformation trend, a weighted correction method is used for adjustment. Specifically, each correction coefficient is multiplied by the predicted deformation value at the corresponding time node and under the corresponding working condition, and then a weighted sum is obtained to obtain the comprehensively corrected deformation. The system calculates the deformation trend prediction value; simultaneously, it retrieves the design deformation limit value and the warning threshold corresponding to different safety levels from the standardized structural design parameter set. The first-level warning threshold is 70% of the design limit, the second-level warning threshold is 85% of the design limit, and the over-limit threshold is 100% of the design limit. The corrected deformation trend prediction value is compared with each threshold one by one to determine whether the deformation state at each time node is in the normal evolution range, the first-level warning range, the second-level warning range, or the over-limit range. Finally, it obtains the predicted deformation trend of the monitoring area, which includes the deformation development direction, the deformation rate change law, the warning threshold trigger time, and the safety risk level. Other methods can be used in other embodiments, which are not limited here.

[0091] It should be noted that the preliminary deformation trend in this application represents the original development trend of temporal deformation in the deformation monitoring area of ​​the target civil engineering structure, reflecting the original change law of deformation characteristic parameters integrating visual and physical dimensions at short-term, medium-term, and long-term time scales; the predicted deformation trend represents the deformation safety development trend in the deformation monitoring area of ​​the target civil engineering structure, reflecting the deformation evolution law and safety risk level of the target civil engineering structure under actual service conditions, covering information such as deformation development direction, early warning threshold trigger time, and whether there is a risk of exceeding limits, which can directly provide decision-making basis for structural safety operation and maintenance and risk management.

[0092] In step 105, an early warning is given for the deformation status of the monitored area in the target civil engineering structure based on the predicted deformation trend.

[0093] In some embodiments, the early warning of the deformation state of a monitored area in a target civil engineering structure based on the predicted deformation trend can be achieved by the following steps:

[0094] Extract various early warning indicators from the predicted deformation trend;

[0095] Establish a three-level early warning threshold system for target civil engineering structures;

[0096] Each early warning indicator is compared with the three-level early warning threshold system to provide early warning of the deformation status of the monitored area in the target civil engineering structure.

[0097] In specific implementation, the extraction of various early warning indicators from the predicted deformation trend can be achieved in the following way: Early warning indicators strongly correlated with the deformation state of the monitoring area are retrieved from the predicted deformation trend. Specifically, these indicators cover the deformation development direction of key stress points in the monitoring area, the deformation rate per unit time, the cumulative deformation increment within the monitoring period, and the predicted time nodes for reaching each level of early warning threshold. Simultaneously, the effectiveness of the extracted indicators is verified. The significance level is set to 0.05 using the Grubbs criterion to eliminate abnormal indicator data caused by model fitting errors or data transmission interference. The indicators are bound to the spatial coordinates of the monitoring points to ensure that each early warning indicator corresponds to a specific component or part in the monitoring area, thus obtaining each early warning indicator. Other methods can also be used in other embodiments, which are not limited here.

[0098] In addition, in specific implementation, the three-level early warning threshold system for the target civil engineering structure can be determined in the following way: based on the structural safety level, design deformation limits of different components, and design bearing capacity of components in the structural design parameter set, a differentiated three-level early warning threshold system is constructed in combination with the load-bearing and non-load-bearing components in the monitoring area. For load-bearing components, the first-level early warning threshold is set at 70% of the design deformation limit, corresponding to a state where the structural deformation is slightly deviating from the normal range and there is no obvious risk; the second-level early warning threshold is set at 85% of the design deformation limit, corresponding to a state where the structural deformation is continuously increasing and approaching the safety threshold; the over-limit early warning threshold is set at 100% of the design deformation limit, corresponding to a dangerous state where the structural deformation exceeds the safe bearing capacity. For non-load-bearing components, the early warning thresholds can be appropriately relaxed, with the first-level early warning threshold at 80% of the design deformation limit and the second-level early warning threshold at 90% of the design deformation limit. At the same time, the applicable scenarios, handling priorities, and responsible entities corresponding to each early warning level are clearly defined, and the three-level early warning threshold system for the target civil engineering structure is finally obtained. Other methods can also be used in other embodiments, which are not limited here.

[0099] It should be noted that the early warning indicators in this application represent a set of quantitative judgment parameters for the monitoring area and safety early warning field of the target civil engineering structure, reflecting the dynamic evolution characteristics, spatial distribution uniformity, and potential safety risk tendency of the structural deformation in the monitoring area; the three-level early warning threshold system represents the graded safety control benchmark of the monitoring area of ​​the target civil engineering structure, reflecting the safety boundary range of the target civil engineering structure from normal service to near-critical and then to dangerous state, providing a clear quantitative basis for the graded judgment and differentiated handling of deformation early warning.

[0100] In addition, in specific implementation, comparing each early warning indicator with the three-level early warning threshold system to provide early warning of the deformation status of the monitored area in the target civil engineering structure can be achieved in the following way: The extracted early warning indicators are compared point-by-point and dimension-by-dimensional with the three-level early warning threshold system. For monitoring areas where the early warning indicators are within the first-level early warning threshold range, a 5-point moving average filtering algorithm is used to perform secondary smoothing and verification of the predicted deformation trend data. Simultaneously, the influence ratio of environmental factors such as temperature, humidity, and instantaneous load on deformation is analyzed to eliminate false early warnings caused by occasional interference, triggering an encrypted monitoring command to increase the monitoring frequency from the original 24 hours / time to 6 hours / time. For areas where the early warning indicators are within the second-level early warning threshold range, historical deformation monitoring data under the same working conditions and seasons are retrieved simultaneously for trend comparison analysis. The driving factors of deformation growth, such as overload, material aging, and foundation defects, are verified using finite element model inversion. Uneven settlement triggers on-site manual inspection and load check instructions, focusing on verifying component connection nodes, anchorage status of embedded parts, and surrounding load distribution, generating and submitting inspection reports. For areas where indicators exceed the over-limit warning threshold, an emergency data verification process is immediately initiated, organizing professional technicians to review the boundary conditions, parameter values, and fitting accuracy of the prediction model. Simultaneously, real-time monitoring data is cross-validated. After confirming the effectiveness of the warning, emergency control and temporary structural reinforcement instructions are triggered, implementing emergency measures such as load limits and temporary supports in the monitored area, evacuating surrounding personnel, and initiating structural safety assessment procedures. Based on the results of the entire process of comparison, verification, and handling, a formal warning report is generated, including the spatial location of the warning area, warning level, risk causes, development trends, and handling recommendations. This completes the accurate and graded warning of the deformation status of the target civil engineering structure monitoring area. Other methods can also be used in other embodiments, which are not limited here.

[0101] In another aspect, in some embodiments, this application provides an automated measurement system for monitoring deformation in civil engineering projects, with reference to... Figure 4 The figure is a schematic diagram of the structure of an automated measurement system for civil engineering deformation monitoring according to some embodiments of this application. The automated measurement system 400 for civil engineering deformation monitoring includes: a data acquisition module 401, a processing module 402, and an execution module 403, which are described below:

[0102] The acquisition module 401 in this application is mainly used to acquire time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure.

[0103] Processing module 402, in this application, is used to extract a sequence of visual feature vectors characterizing the structural deformation of the monitoring area from the time-series visual data, and to extract a sequence of physical feature vectors from the time-series physical quantity data;

[0104] It should be noted that the processing module 402 in this application is also used to determine the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, and to cross-fuse the visual feature vector sequence and the physical feature vector sequence through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area.

[0105] In addition, it should be noted that the processing module 402 in this application is also used to obtain the structural design parameter set of the target civil engineering structure, and to predict and analyze the deformation trend of the monitoring area based on the structural design parameter set and the fused deformation feature sequence, so as to obtain the predicted deformation trend of the monitoring area.

[0106] The execution module 403 in this application is mainly used to provide early warning of the deformation status of the monitoring area in the target civil engineering structure based on the predicted deformation trend.

[0107] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described automated measurement method for deformation monitoring in civil engineering.

[0108] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for implementing an automated measurement method for deformation monitoring in civil engineering, according to some embodiments of this application. The automated measurement method for deformation monitoring in civil engineering described in the above embodiments can be implemented through... Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0109] Processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0110] The communication bus 502 can be used to transmit information between the aforementioned components.

[0111] Memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. Memory 503 may exist independently and be connected to processor 501 via communication bus 502. Memory 503 may also be integrated with processor 501.

[0112] The memory 503 stores program code for executing the scheme of this application, and its execution is controlled by the processor 501. The processor 501 executes the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0113] Communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0114] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0115] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0116] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described automated measurement method for deformation monitoring in civil engineering.

[0117] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0118] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. An automated measurement method for deformation monitoring in civil engineering, characterized in that, Includes the following steps: Collect time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure; A sequence of visual feature vectors characterizing the structural deformation of the monitoring area is extracted from the time-series visual data, and a sequence of physical feature vectors is extracted from the time-series physical quantity data; Based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined. The visual feature vector sequence and the physical feature vector sequence are cross-fused through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area. Obtain the set of structural design parameters for the target civil engineering structure, and predict and analyze the deformation trend of the monitored area based on the set of structural design parameters and the fused deformation feature sequence to obtain the predicted deformation trend of the monitored area. Based on the predicted deformation trend, an early warning is issued for the deformation status of the monitored area in the target civil engineering structure.

2. The method as described in claim 1, characterized in that, Extracting the sequence of visual feature vectors characterizing the structural deformation of the monitored area from the temporal visual data specifically includes: The temporal visual data is preprocessed to obtain preprocessed temporal visual data; Deformation features are extracted from the preprocessed temporal visual data to obtain a sequence of visual feature vectors characterizing the structural deformation of the monitored area.

3. The method as described in claim 1, characterized in that, Extracting the physical feature vector sequence from the time-series physical quantity data specifically includes: The time-series physical quantity data is preprocessed to obtain preprocessed time-series physical quantity data; Physical quantity features are extracted from the preprocessed time-series physical quantity data to obtain a sequence of physical feature vectors.

4. The method as described in claim 1, characterized in that, Based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined, specifically including: Acquire historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure; Determine the correlation coefficients between the parameters in the historical monitoring data; The coupling relationship between deformation and visual and physical quantities in the target civil engineering structure is determined based on all correlation coefficients and the historical monitoring data.

5. The method as described in claim 1, characterized in that, The visual feature vector sequence and the physical feature vector sequence are cross-fused through the coupling relationship to obtain the fused deformation feature sequence characterizing the monitoring area, specifically including: The visual feature vector sequence and the physical feature vector sequence are time-aligned to obtain the aligned visual feature vector sequence and physical feature vector sequence; Based on the aforementioned coupling relationship, the aligned visual feature vector sequence and physical feature vector sequence are weighted and fused according to their corresponding dimensions to obtain fused feature vectors for each dimension; All the fused feature vectors are cross-stacked to obtain a fused deformation feature sequence characterizing the monitored area.

6. The method as described in claim 1, characterized in that, Based on the structural design parameter set and the fused deformation feature sequence, the deformation trend of the monitored area is predicted and analyzed to obtain the predicted deformation trend of the monitored area, which specifically includes: The structural design parameter set and the fused deformation feature sequence are standardized to obtain a standardized structural design parameter set and fused deformation feature sequence. The standardized fused deformation feature sequence is fitted with deformation trends at multiple time scales to obtain the preliminary deformation trend of the monitored area. The preliminary deformation trend is corrected based on the standardized set of structural design parameters to obtain the predicted deformation trend of the monitored area.

7. The method as described in claim 1, characterized in that, The specific methods for providing early warning of the deformation status of the monitored area in the target civil engineering structure based on the predicted deformation trend include: Extract various early warning indicators from the predicted deformation trend; Establish a three-level early warning threshold system for target civil engineering structures; Each early warning indicator is compared with the three-level early warning threshold system to provide early warning of the deformation status of the monitored area in the target civil engineering structure.

8. An automated measurement system for deformation monitoring in civil engineering, characterized in that, include: The acquisition module is used to acquire time-series visual data and time-series physical quantity data of the monitoring area in the target civil engineering structure; The processing module is used to extract a sequence of visual feature vectors characterizing the structural deformation of the monitoring area from the time-series visual data, and to extract a sequence of physical feature vectors from the time-series physical quantity data. The processing module is also used to determine the coupling relationship between deformation and visual and physical quantities in the target civil engineering structure based on historical monitoring data of visual and physical quantities related to deformation in the target civil engineering structure, and to cross-fuse the visual feature vector sequence and the physical feature vector sequence through the coupling relationship to obtain a fused deformation feature sequence characterizing the monitoring area. The processing module is also used to obtain the structural design parameter set of the target civil engineering structure, and to predict and analyze the deformation trend of the monitoring area based on the structural design parameter set and the fused deformation feature sequence, so as to obtain the predicted deformation trend of the monitoring area. The execution module is used to provide early warning of the deformation status of the monitored area in the target civil engineering structure based on the predicted deformation trend.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the automated measurement method for deformation monitoring in civil engineering as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the automated measurement method for deformation monitoring in civil engineering as described in any one of claims 1 to 7.

Images