A deformation monitoring and early warning method and system for existing tunnels

Abstract

The present invention relates to the field of tunnel early warning, and specifically to a deformation monitoring and early warning method and system for existing tunnels. By extracting a historical monitoring image sequence of multiple image monitoring points of a reference tunnel and analyzing the historical monitoring image sequence, a target time point when the reference tunnel is deformed is obtained. Then, combined with monitoring data of various geological structure categories of the reference tunnel, the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency of the tunnel lining in a period of time before the deformation occurs are extracted to construct deformation feature template data. When deformation monitoring is performed on an existing tunnel, deformation feature data of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency in the existing tunnel are extracted, and the deformation feature data and the deformation feature template data are compared for similarity to determine whether there is a deformation risk.

Description

Technical Field

[0001] The present invention relates to the field of tunnel early warning, and in particular to a deformation monitoring and early warning method and system for an existing tunnel. Background Art

[0002] During tunnel construction and operation, deformation monitoring and early warning are essential. Tunnel deformation early warning technology can monitor tunnel structure deformation in real time and issue alerts if abnormal changes are detected. This helps prevent or mitigate personal injury and property damage caused by disasters such as tunnel collapse and crack expansion.

[0003] Existing technologies for monitoring deformation in existing tunnels primarily rely on measuring the tunnel's inner wall with a rangefinder or using contact monitoring equipment to detect subtle deformations on the inner wall and provide early warnings. However, monitoring only the tunnel lining surface fails to provide stress data and trends in the surrounding rock. Therefore, it's impossible to predict lining changes based on stress changes in advance, making it impossible to reinforce the lining in advance to prevent significant deformation. Summary of the Invention

[0004] In view of this, an object of the present invention is to provide a deformation monitoring and early warning method and system for an existing tunnel to solve the above technical problems.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A deformation monitoring and early warning method for an existing tunnel according to the present invention comprises the following steps:

[0007] Acquire a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points;

[0008] determining a target location where lining deformation occurs in the existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence;

[0009] Classifying multiple mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data of a target time point and multiple historical time points of a target duration before the target time point;

[0010] Acquire monitoring data sequences and geological structure categories of a target location of an existing tunnel at multiple time points within a target time period before a current time point, and extract deformation feature data from the monitoring data sequences;

[0011] The target deformation feature template data of the target position is determined based on the geological structure category, the deformation feature data is compared with the target deformation feature template data for similarity, and deformation warning is performed on the existing tunnel based on the comparison result.

[0012] In one embodiment of the present application, determining a target location where lining deformation occurs in an existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence includes:

[0013] Dividing the historical monitoring images of any two adjacent time points in the historical monitoring image sequence into the same control group;

[0014] The first frame of image in the control group is used as a reference image, and the second frame of image is used as a control image;

[0015] Preprocessing the reference image and the control image respectively to obtain a first preprocessed image and a second preprocessed image, wherein the preprocessing method includes grayscale conversion and high-pass filtering;

[0016] Adjusting the brightness of the first pre-processed image and the second pre-processed image to obtain a first image and a second image with consistent brightness; and extracting candidate gap regions having grayscale values ​​less than a preset threshold from the first image and the second image;

[0017] Extracting contour features of the gap candidate area; performing morphological processing on the contour features of the gap candidate area to obtain closed contour features of the gap candidate area; calculating the size and roundness of the closed contour features, and using contour features that meet a preset size range and roundness range as the top arch gap contour;

[0018] Drawing the closed contour features of the first image and the closed contour features of the second image into a blank image respectively to obtain a first contour image and a second contour image;

[0019] Extracting the centroid of the top arch gap contour and other closed contour features, and registering the first contour image and the second contour image based on the centroid of the top arch gap contour;

[0020] The closed contour features with the closest centroid distance between the first contour image and the second contour image are used as reference contours, and the overlap rate ρ of each set of reference contours is calculated. The mathematical expression of the overlap rate ρ is:

[0021]

[0022] Where A ol is the area of ​​the overlapping part of the control contour, A max is the area of ​​the larger contour among the control contours;

[0023] The control contour with an overlap rate ρ less than a preset threshold is restored to the control image to obtain a deformed position; and the time point of the control image is used as the target time point.

[0024] In one embodiment of the present application, brightness adjustment is performed on the first pre-processed image and the second pre-processed image to obtain a first image and a second image with consistent brightness, including:

[0025] Normalizing the grayscale values ​​of all pixels in the first preprocessed image and the second preprocessed image to obtain a first normalized image and a second normalized image, respectively;

[0026] The first normalized image and the second normalized image are reassigned using the same grayscale range to obtain a first image and a second image with consistent brightness.

[0027] In one embodiment of the present application, deformation feature template data is constructed based on target monitoring data of mechanical monitoring locations of each geological structure category, including:

[0028] Calculating the variance of each monitoring parameter at multiple historical time points in the target monitoring data;

[0029] When there is a target monitoring parameter with a variance greater than a preset corresponding variance threshold in the target monitoring data, extracting an abnormal value of the target monitoring parameter, calculating an average value of other monitoring parameters at multiple historical time points, and constructing a feature vector based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and a first parameter flag bit, wherein the value of the first parameter flag bit indicates that the feature vector is a mutation feature;

[0030] When there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, calculating the average value of each monitoring parameter at multiple historical time points, and constructing a feature vector based on the average value of each monitoring parameter at multiple historical time points and a second flag bit, wherein the second parameter flag bit indicates that the feature vector is a persistent feature;

[0031] Perform density clustering on all feature vectors to obtain multiple feature clusters;

[0032] The feature cluster whose number of feature vectors is greater than a preset number threshold is taken as the target feature cluster, and the average value of each monitoring parameter in the target feature cluster is calculated to obtain deformation feature template data.

[0033] In one embodiment of the present application, extracting deformation feature data from the monitoring data sequence includes:

[0034] Calculating the variance of multiple monitoring parameters in the monitoring data sequence at multiple time points;

[0035] When a target monitoring parameter having a variance greater than a preset corresponding variance threshold exists in the monitoring data sequence, an abnormal value of the target monitoring parameter is extracted, and an average value of other monitoring parameters at multiple historical time points is calculated, and a feature vector is constructed based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and a first parameter flag, and the feature vector is used as deformation feature data;

[0036] When there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, the average value of each monitoring parameter at multiple historical time points is calculated, and deformation feature data is constructed based on the average value of each monitoring parameter at multiple historical time points and the second flag.

[0037] In one embodiment of the present application, an abnormal value of a target monitoring parameter is extracted, and the average value of other monitoring parameters at multiple historical time points is calculated. A feature vector is constructed based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and the first parameter flag, including:

[0038] Calculate the average value of the target monitoring parameter at multiple historical time points; and calculate the deviation rate of each value of the target monitoring parameter from the average value; take the value with a deviation rate greater than the set deviation ratio as an outlier, cluster all outliers to obtain multiple outlier clusters; calculate the average value of each outlier cluster to obtain the outlier value of the target monitoring parameter;

[0039] Each abnormal value of the target monitoring parameter is combined with the average value and the first flag of other monitoring parameters at multiple historical time points to obtain one or more feature vectors.

[0040] In one embodiment of the present application, determining target deformation feature template data of the target location based on the geological structure category, and performing similarity comparison between the deformation feature data and the target deformation feature template data include:

[0041] Determine target deformation feature template data having the same geological structure category and the same marker position as the target location;

[0042] Calculate the deviation rate D of each monitoring parameter in the deformation feature data and the target deformation feature template data, wherein the deviation rate D i The mathematical expression is:

[0043]

[0044] Where x i is the value of the i-th monitoring parameter in the deformation feature data, x′ i is the value of the i-th monitoring parameter in the target deformation feature template data, x max is x i and x′ i The maximum value in ;

[0045] The deviation rate D at each position i When both are smaller than a preset deviation rate threshold, it is determined that the deformation feature data is similar to the target deformation feature template data; otherwise, it is determined that the deformation feature data is not similar to the target deformation feature template data.

[0046] In one embodiment of the present application, deformation warning of an existing tunnel is performed based on the comparison result, including:

[0047] When the comparison result is similar, the monitoring data sequence is marked as having deformation risk and a warning message is generated, and the warning message is sent to the target object; when the comparison result is dissimilar, the monitoring data sequence is marked as not having deformation risk.

[0048] In one embodiment of the present application, the target location is a pre-marked location where strain is likely to occur.

[0049] This application also provides a deformation monitoring and early warning system for an existing tunnel, comprising:

[0050] a sample acquisition module, configured to acquire a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points;

[0051] a deformation event screening module for determining a target location where lining deformation occurs in an existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence;

[0052] a template feature extraction module for classifying a plurality of mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data at a target time point and at multiple historical time points of a target duration before the target time point;

[0053] A feature extraction module is used to obtain a monitoring data sequence and a geological structure category of a target position of an existing tunnel at multiple time points within a target time period before a current time point, and to extract deformation feature data from the monitoring data sequence;

[0054] The comparison and warning module is used to determine the target deformation feature template data of the target location based on the geological structure category, perform similarity comparison between the deformation feature data and the target deformation feature template data, and perform deformation warning for the existing tunnel based on the comparison result.

[0055] The beneficial effects of the present invention are as follows: a deformation monitoring and early warning method and system of an existing tunnel of the present invention extracts a historical monitoring image sequence of multiple image monitoring points of a reference tunnel and analyzes the historical monitoring image sequence to obtain the target time point when the reference tunnel is deformed. Then, the monitoring data of various geological structure categories of the reference tunnel are combined to extract the characteristics of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency of the tunnel lining in a period of time before the deformation occurs, thereby constructing deformation feature template data. When deformation monitoring is performed on an existing tunnel, deformation feature data of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency in the existing tunnel are extracted, and the deformation feature data and the deformation feature template data are compared for similarity to determine whether there is a deformation risk. This application provides an early warning method different from lining surface detection by comprehensively analyzing the various factors that affect tunnel lining deformation. Combined with the existing lining surface detection technology, deformation monitoring and early warning of tunnel deformation can be performed in all directions. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0057] Figure 1 is a cross-sectional structural diagram of an existing tunnel shown in an embodiment of the present application;

[0058] Figure 2 This is a diagram illustrating an application scenario of a deformation monitoring and early warning method for an existing tunnel in an embodiment of the present application;

[0059] Figure 3 This is a flow chart of a deformation monitoring and early warning method for an existing tunnel shown in one embodiment of the present application;

[0060] Figure 4 Schematic diagram of a deformation determination process based on machine vision in one embodiment of the present application;

[0061] Figure 5 This is a structural diagram of a deformation monitoring and early warning system for an existing tunnel shown in one embodiment of the present application. DETAILED DESCRIPTION

[0062] The following describes the embodiments of the present invention through specific examples. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention.

[0063] Figure 1 This is a cross-sectional structural diagram of an existing tunnel shown in an embodiment of the present application. Figure 1 As shown, existing tunnels include primary support formed during the initial construction phase, and segments 110, which are structurally reinforced externally. The primary support and segments 110 form a complete lining structure, with surrounding rock surrounding the lining. Contact pressure is generated at the interface between the surrounding rock and the lining. This contact pressure is not constant and can vary due to factors such as temperature and humidity, groundwater activity, and vibrations from vehicles traveling through the tunnel. Large variations can easily cause deformation of the lining.

[0064] Figure 2 This is an application scenario diagram of a deformation monitoring and early warning method for an existing tunnel shown in an embodiment of the present application, such as Figure 2 As shown, the present application sets a pressure sensor 210 between the surrounding rock and the lining, and sets a temperature and humidity sensor 220, a vibration sensor 230, etc. in the internal structure of the surrounding rock. The above sensors are connected to the host through wires for analysis.

[0065] In addition, a camera 240 is installed in the tunnel. The camera 240 captures images of the tunnel vault and cracks and sends them to the host computer for visual analysis. Combined with internal structural parameter analysis and surface structure visual analysis, comprehensive deformation monitoring of the tunnel can be performed. In this application, visual analysis can use other existing visual analysis technologies or the visual analysis solution described below.

[0066] Figure 3 This is a flow chart of a deformation monitoring and early warning method for an existing tunnel shown in one embodiment of the present application. Figure 3 As shown in the figure, a deformation monitoring and early warning method for an existing tunnel according to this embodiment may include steps S310 to S350:

[0067] S310, acquiring a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points;

[0068] In the present application, it is easy to understand that the historical monitoring data sequence is regularly collected and constructed by sensors pre-buried in the tunnel structure, and the historical monitoring image sequence is regularly collected and acquired by fixed cameras set up in the tunnel.

[0069] Geological structure data is obtained from survey data collected during tunnel construction. Examples include bedding, joints, faults, folds, and other structures. Different geological structures have different effects on the lining, so this application analyzes each structure separately.

[0070] In addition, the reference tunnel in this application can be the historical data of a tunnel or the historical data of multiple tunnels of the same category, and there is no limitation here.

[0071] S320: Determine a target location where lining deformation occurs in the existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence; the target location is a pre-marked location where strain is likely to occur, such as a gap, a crown, etc.

[0072] Figure 4 FIG. 1 is a flow chart of deformation determination based on machine vision in an embodiment of the present application, as shown in FIG. Figure 4 As shown, in this application, machine vision technology is used to determine the presence of deformation in historical monitoring images, and then determine the time point when the tunnel lining deformation occurs, including:

[0073] S321, dividing the historical monitoring images of any two adjacent time points in the historical monitoring image sequence into the same control group; and pairing the historical monitoring images of adjacent time points into groups for subsequent comparative analysis.

[0074] S322, using the previous frame image in the control group as a reference image and the next frame image as a control image;

[0075] S323, preprocessing the reference image and the control image respectively to obtain a first preprocessed image and a second preprocessed image, wherein the preprocessing method includes grayscale conversion and high-pass filtering;

[0076] The image undergoes pre-processing steps such as grayscale conversion and high-pass filtering to highlight edges and reduce unnecessary details. This improves the performance of subsequent processing steps, making important features such as cracks more visible.

[0077] S324: Perform brightness adjustment on the first pre-processed image and the second pre-processed image to obtain a first image and a second image with consistent brightness; and extract candidate gap regions having grayscale values ​​less than a preset threshold from the first image and the second image.

[0078] Adjust the brightness of the two images to make them consistent, and filter out areas that may represent cracks based on grayscale values. This ensures that images taken at different time points are compared under the same conditions, improving the accuracy of crack detection. Specifically, the brightness adjustment process includes:

[0079] S3241: Normalize the grayscale values ​​of all pixels in the first preprocessed image and the second preprocessed image to obtain a first normalized image and a second normalized image, respectively;

[0080] The mathematical expression of the normalized image is:

[0081]

[0082] Where Normalized(i,j) represents the value of pixel (i,j) in the normalized image, gray(i,j) is the grayscale value of pixel (i,j) in the preprocessed image, and gray min is the minimum gray value in the preprocessed image, gray max is the maximum grayscale value in the preprocessed image.

[0083] S3242: Reassign the first normalized image and the second normalized image using the same grayscale range to obtain a first image and a second image with consistent brightness.

[0084] This application expands the value range of the first normalized image and the second normalized image to 0-255, so the grayscale values ​​of the first image and the second image are obtained in the following manner.

[0085]

[0086] The image processed by the above process has consistent brightness, so that when grayscale features are subsequently extracted, different grayscale features can accurately represent different areas, such as gaps.

[0087] In the image, since the gap is presented as a shadow, the grayscale value is small. This application sets a grayscale threshold to filter out the gap area.

[0088] S325, extracting contour features of the gap candidate region; performing morphological processing on the contour features of the gap candidate region to obtain closed contour features of the gap candidate region; calculating the size and roundness of the closed contour features, and using contour features that meet a preset size range and roundness range as the arch gap contour;

[0089] After obtaining the gap candidate regions with lower grayscale values, in order to avoid the influence of patterns on the wall, some wall tiles or ground cracks, the present application screens the gap candidate regions based on contour features.

[0090] First, the Canny operator is used to extract the contour features of the gap candidate area. These contour features are then eroded and expanded to form complete and closed contour features. Geometric information, such as size and roundness, is then extracted from the closed contours to identify the contours of the arch gap and wall gap that meet the preset size range and roundness.

[0091] like Figure 2 As shown in the figure, there are connected arch gaps and wall gaps at the lining connection. These gaps will change significantly when the lining deforms. Therefore, the contours of these gaps are used as the reference parameters for whether the lining has deformed.

[0092] S326, drawing the closed contour features of the first image and the closed contour features of the second image into a blank image respectively, to obtain a first contour image and a second contour image;

[0093] The processed closed contour features are mapped onto a blank image to generate a first contour image and a second contour image containing only contour information. This simplifies the image content, allowing subsequent processes to focus on comparing cracks and other key features.

[0094] S327, extracting the centroid of the top arch gap contour and other closed contour features, and registering the first contour image and the second contour image based on the centroid of the top arch gap contour;

[0095] There are often multiple roof arch gap contours in one image, so by extracting the centroids of the multiple roof arch gap contours, the first contour image and the second contour image can be accurately aligned.

[0096] S328: The closed contour features with the closest centroid distance between the first contour image and the second contour image are used as reference contours, and the overlap rate ρ of each set of reference contours is calculated. The mathematical expression of the overlap rate ρ is:

[0097]

[0098] Where A ol is the area of ​​the overlapping part of the control contour, Amax is the area of ​​the larger contour among the control contours;

[0099] After registration, the centroids of each contour are used to combine them, and the closest gap contour is used as the gap contour at the same location but at a different time. The area of ​​the overlapping portion of the two contours and the overlap ratio are then calculated. The overlap ratio is used to determine whether the gap has changed significantly.

[0100] S329, restoring the control contour with a coincidence rate ρ less than a preset threshold to the control image to obtain a deformed position; and using the time point of the control image as the target time point.

[0101] If the overlap ratio ρ is less than a preset threshold, it means the two images cannot be accurately aligned, indicating a significant change in the gap. The deformation of the lining at the gap can then be determined. The reference contour is then restored to the reference image to determine the location of the deformation. The time the reference image was captured is used as the target time of the deformation.

[0102] S330, classifying multiple mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data at a target time point and multiple historical time points of a target duration before the target time point;

[0103] After the above-mentioned machine vision analysis, the target time point when the deformation occurs is obtained. The monitoring data can be analyzed based on the target time point to obtain the characteristics of the monitoring data when the lining deformation occurs and in the period before the lining deformation.

[0104] In this example, lining deformation may be caused by a combination of long-term, stable conditions, such as a long-term high temperature and high humidity environment that results in a high lining load. It may also be caused by a short-term, sudden change in monitoring parameters, such as a sudden rainstorm that causes a high water content in the surrounding rock, which in turn increases groundwater pressure and changes in surrounding rock properties, leading to lining deformation.

[0105] In view of this, this application analyzes not only the target time point, but also the monitoring data for a period of time before the target time point, specifically including:

[0106] S331, calculating the variance of each monitoring parameter at multiple historical time points in the target monitoring data;

[0107] Variance can reflect the volatility of the data. If the variance is large, it means that the values ​​of the monitoring parameters have changed dramatically within a period of time between the target time points. In this case, it is speculated that the deformation is caused by the drastic change in the monitoring parameters. If the variance is small, it means that the deformation is not caused by drastic changes in the values ​​of the monitoring parameters, but rather by the values ​​of multiple monitoring parameters remaining within a stable range for a long time.

[0108] S332, when there is a target monitoring parameter in the target monitoring data whose variance is greater than a preset corresponding variance threshold, extracting an abnormal value of the target monitoring parameter, calculating an average value of other monitoring parameters at multiple historical time points, and constructing a feature vector based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and a first parameter flag, wherein the value of the first parameter flag indicates that the feature vector is a mutation feature;

[0109] If there is a target monitoring parameter in the target monitoring data whose variance is greater than the preset corresponding variance threshold, the data is regarded as data deformed by a sudden abnormal value, and a feature vector is constructed, which includes a first flag bit.

[0110] The eigenvector can be: (x1, x2, x3, x4, x5, x6, m), where x1 is the load ratio, x2 is the temperature, x3 is the humidity, x4 is the water content of the surrounding rock, x5 is the vibration amplitude, x6 is the vibration frequency, and m is the flag bit.

[0111] For target monitoring data of mutation type, this application uses the following methods to construct feature vectors, including:

[0112] S3321, calculating the average value of the target monitoring parameter at multiple historical time points; and calculating the deviation rate of each value of the target monitoring parameter from the average value; treating values ​​with a deviation rate greater than a set deviation ratio as outliers, clustering all outliers to obtain multiple outlier clusters; calculating the average value of each outlier cluster to obtain the outlier value of the target monitoring parameter;

[0113] In this embodiment, the deviation rate is used to screen outliers. Outliers may be clustered in one interval or in multiple intervals. It may be necessary to cluster all outliers to obtain clustering features of the outliers.

[0114] The clusters obtained by clustering may be one or more, so the typical outliers obtained may also be one or more.

[0115] S3322: Combine each abnormal value of the target monitoring parameter with the average values ​​and first flags of other monitoring parameters at multiple historical time points to obtain one or more feature vectors.

[0116] If there are multiple typical outliers, a feature vector can be constructed based on the outlier and the average values ​​and first flags of other monitoring parameters at multiple historical time points.

[0117] S333: When there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, calculating the average value of each monitoring parameter at multiple historical time points, and constructing a feature vector based on the average value of each monitoring parameter at multiple historical time points and a second flag bit, wherein the second parameter flag bit indicates that the feature vector is a persistent feature;

[0118] If the lining deformation is not caused by a sudden change, the average value of the monitoring parameters during this period is taken as its eigenvalue to construct the eigenvector.

[0119] S334, performing density clustering on all feature vectors to obtain multiple feature clusters;

[0120] In this application, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm is used to perform density clustering, thereby obtaining clusters of multiple feature vectors.

[0121] S335 , taking a feature cluster whose number of feature vectors is greater than a preset threshold as a target feature cluster, and calculating an average value of each monitoring parameter in the target feature cluster to obtain deformation feature template data.

[0122] If the amount of data in a cluster is large, it means that a large number of feature vectors are concentrated in one cluster. The feature vectors in this cluster can reflect a feature template corresponding to lining deformation. Therefore, the value of each monitoring parameter in the cluster is averaged to serve as the typical value of the deformation feature template data.

[0123] S340, obtaining a monitoring data sequence and a geological structure category of a target location of an existing tunnel at multiple time points within a target time period before the current time point, and extracting deformation feature data from the monitoring data sequence;

[0124] The existing tunnel is the target tunnel for deformation monitoring in this application. The sensors described above are used to collect monitoring data sequences at multiple time points within the target duration before the current time point, and then feature extraction is performed, including:

[0125] S341, calculating the variance of multiple monitoring parameters in the monitoring data sequence at multiple time points;

[0126] S342, when there is a target monitoring parameter in the monitoring data sequence whose variance is greater than a preset corresponding variance threshold, extracting an abnormal value of the target monitoring parameter, calculating an average value of other monitoring parameters at multiple historical time points, and constructing a feature vector based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and the first parameter flag, and using the feature vector as deformation feature data;

[0127] S343, when there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, calculate the average value of each monitoring parameter at multiple historical time points, and construct deformation feature data based on the average value of each monitoring parameter at multiple historical time points and the second flag.

[0128] S350: determining target deformation feature template data of the target location based on the geological structure category, performing a similarity comparison between the deformation feature data and the target deformation feature template data, and performing deformation warning for the existing tunnel based on the comparison result.

[0129] Finally, the pre-built target deformation feature template data is used for comparison to determine whether there is a lining risk at present, including:

[0130] S351, determining target deformation feature template data having the same geological structure category and the same marker position as the target location;

[0131] S352, calculating the deviation rate D of each monitoring parameter in the deformation feature data and the target deformation feature template data, wherein the deviation rate D i The mathematical expression is:

[0132]

[0133] Where x i is the value of the i-th monitoring parameter in the deformation feature data, x′ i is the value of the i-th monitoring parameter in the target deformation feature template data, x max is x i and x′ i The maximum value in ;

[0134] S353, deviation rate D at each position i When both are smaller than a preset deviation rate threshold, it is determined that the deformation feature data is similar to the target deformation feature template data; otherwise, it is determined that the deformation feature data is not similar to the target deformation feature template data.

[0135] In this embodiment, the deviation rate is used to evaluate the similarity between the two. If the comparison result is similar, the monitoring data sequence is marked as having a deformation risk and a warning message is generated and sent to the target object. If the comparison result is dissimilar, the monitoring data sequence is marked as not having a deformation risk.

[0136] The present invention provides a deformation monitoring and early warning method for an existing tunnel. The method extracts a historical monitoring image sequence of multiple image monitoring points of a reference tunnel and analyzes the historical monitoring image sequence to obtain a target time point when the reference tunnel is deformed. Then, the monitoring data of various geological structure categories of the reference tunnel are combined to extract the characteristics of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency of the tunnel lining in a period of time before the deformation occurs, thereby constructing deformation feature template data. When deformation monitoring is performed on an existing tunnel, deformation feature data of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency in the existing tunnel are extracted, and the deformation feature data and the deformation feature template data are compared for similarity to determine whether there is a deformation risk. The present application provides an early warning method that is different from lining surface detection by comprehensively analyzing the various factors that affect tunnel lining deformation. Combined with existing lining surface detection technology, deformation monitoring and early warning of tunnel deformation can be performed in all directions.

[0137] like Figure 5 As shown, the present application also provides a deformation monitoring and early warning system for an existing tunnel, comprising:

[0138] a sample acquisition module, configured to acquire a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points;

[0139] a deformation event screening module for determining a target location where lining deformation occurs in an existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence;

[0140] a template feature extraction module for classifying a plurality of mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data at a target time point and at multiple historical time points of a target duration before the target time point;

[0141] A feature extraction module is used to obtain a monitoring data sequence and a geological structure category of a target position of an existing tunnel at multiple time points within a target time period before a current time point, and to extract deformation feature data from the monitoring data sequence;

[0142] The comparison and warning module is used to determine the target deformation feature template data of the target location based on the geological structure category, perform similarity comparison between the deformation feature data and the target deformation feature template data, and perform deformation warning for the existing tunnel based on the comparison result.

[0143] The present invention provides a deformation monitoring and early warning system for an existing tunnel. The system extracts a historical monitoring image sequence of multiple image monitoring points of a reference tunnel and analyzes the historical monitoring image sequence to obtain a target time point when the reference tunnel is deformed. Then, the monitoring data of various geological structure categories of the reference tunnel are combined to extract the characteristics of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency of the tunnel lining in a period of time before the deformation occurs, thereby constructing deformation feature template data. When deformation monitoring is performed on an existing tunnel, deformation feature data of the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude and vibration frequency in the existing tunnel are extracted, and the deformation feature data and the deformation feature template data are compared for similarity to determine whether there is a deformation risk. The present application provides an early warning method that is different from lining surface detection by comprehensively analyzing the various factors that affect tunnel lining deformation. Combined with existing lining surface detection technology, deformation monitoring and early warning of tunnel deformation can be performed in all directions.

[0144] The above embodiments are only preferred embodiments for fully illustrating the present application, and the protection scope of the present application is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art based on the present application are within the protection scope of the present application.

Claims

1. A deformation monitoring and early warning method for an existing tunnel, characterized in that: Including steps: Acquire a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points; determining a target location where lining deformation occurs in the existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence; Classifying multiple mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data of a target time point and multiple historical time points of a target duration before the target time point; Acquire monitoring data sequences and geological structure categories of a target location of an existing tunnel at multiple time points within a target time period before a current time point, and extract deformation feature data from the monitoring data sequences; The target deformation feature template data of the target position is determined based on the geological structure category, the deformation feature data is compared with the target deformation feature template data for similarity, and deformation warning is performed on the existing tunnel based on the comparison result.

2. The deformation monitoring and early warning method for an existing tunnel according to claim 1 is characterized in that: Determining a target location where lining deformation occurs in an existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence includes: Dividing the historical monitoring images of any two adjacent time points in the historical monitoring image sequence into the same control group; The first frame of image in the control group is used as a reference image, and the second frame of image is used as a control image; Preprocessing the reference image and the control image respectively to obtain a first preprocessed image and a second preprocessed image, wherein the preprocessing method includes grayscale conversion and high-pass filtering; Adjusting the brightness of the first pre-processed image and the second pre-processed image to obtain a first image and a second image with consistent brightness; and extracting candidate gap regions having grayscale values ​​less than a preset threshold from the first image and the second image; Extracting contour features of the gap candidate area; performing morphological processing on the contour features of the gap candidate area to obtain closed contour features of the gap candidate area; calculating the size and roundness of the closed contour features, and using contour features that meet a preset size range and roundness range as the top arch gap contour; Drawing the closed contour features of the first image and the closed contour features of the second image into a blank image respectively to obtain a first contour image and a second contour image; Extracting the centroid of the top arch gap contour and other closed contour features, and registering the first contour image and the second contour image based on the centroid of the top arch gap contour; The closed contour features with the closest centroid distance between the first contour image and the second contour image are used as the reference contours, and the overlap rate of each set of reference contours is calculated. , the overlap rate The mathematical expression is: Where, is the area of ​​the overlapping part of the control contour, is the area of ​​the largest contour among the control contours; The overlap rate The control contour smaller than a preset threshold is restored to the control image to obtain the deformed position; and the time point of the control image is used as the target time point.

3. The deformation monitoring and early warning method for an existing tunnel according to claim 2 is characterized in that: Performing brightness adjustment on the first preprocessed image and the second preprocessed image to obtain a first image and a second image with consistent brightness includes: Normalizing the grayscale values ​​of all pixels in the first preprocessed image and the second preprocessed image to obtain a first normalized image and a second normalized image, respectively; The first normalized image and the second normalized image are reassigned using the same grayscale range to obtain a first image and a second image with consistent brightness.

4. The deformation monitoring and early warning method for an existing tunnel according to claim 1, characterized in that: Deformation feature template data is constructed based on the target monitoring data of the mechanical monitoring location for each geological structure category, including: Calculating the variance of each monitoring parameter at multiple historical time points in the target monitoring data; When there is a target monitoring parameter with a variance greater than a preset corresponding variance threshold in the target monitoring data, extracting an abnormal value of the target monitoring parameter, calculating an average value of other monitoring parameters at multiple historical time points, and constructing a feature vector based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and a first parameter flag bit, wherein the value of the first parameter flag bit indicates that the feature vector is a mutation feature; When there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, calculating the average value of each monitoring parameter at multiple historical time points, and constructing a feature vector based on the average value of each monitoring parameter at multiple historical time points and a second parameter flag, wherein the second parameter flag indicates that the feature vector is a persistent feature; Perform density clustering on all feature vectors to obtain multiple feature clusters; The feature cluster whose number of feature vectors is greater than a preset number threshold is taken as the target feature cluster, and the average value of each monitoring parameter in the target feature cluster is calculated to obtain deformation feature template data.

5. The deformation monitoring and early warning method for an existing tunnel according to claim 1 is characterized in that: Extracting deformation feature data from the monitoring data sequence includes: Calculating the variance of multiple monitoring parameters in the monitoring data sequence at multiple time points; When a target monitoring parameter having a variance greater than a preset corresponding variance threshold exists in the monitoring data sequence, an abnormal value of the target monitoring parameter is extracted, and an average value of other monitoring parameters at multiple historical time points is calculated, and a feature vector is constructed based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and a first parameter flag, and the feature vector is used as deformation feature data; When there is no monitoring parameter with a variance greater than a preset variance threshold in the target monitoring data, the average value of each monitoring parameter at multiple historical time points is calculated, and deformation feature data is constructed based on the average value of each monitoring parameter at multiple historical time points and the second flag.

6. A deformation monitoring and early warning method for an existing tunnel according to claim 4 or 5, characterized in that: Extract the abnormal value of the target monitoring parameter, calculate the average value of other monitoring parameters at multiple historical time points, and construct a feature vector based on the abnormal value of the target monitoring parameter, the average value of other monitoring parameters at multiple historical time points, and the first parameter flag, including: Calculate the average value of the target monitoring parameter at multiple historical time points; and calculate the deviation rate of each value of the target monitoring parameter from the average value; take the value with a deviation rate greater than the set deviation ratio as an outlier, cluster all outliers to obtain multiple outlier clusters; calculate the average value of each outlier cluster to obtain the outlier value of the target monitoring parameter; Each abnormal value of the target monitoring parameter is combined with the average value and the first flag of other monitoring parameters at multiple historical time points to obtain one or more feature vectors.

7. The deformation monitoring and early warning method for an existing tunnel according to claim 1 is characterized in that: Determining target deformation feature template data of the target location based on the geological structure category, and performing similarity comparison between the deformation feature data and the target deformation feature template data, including: Determine target deformation feature template data having the same geological structure category and the same marker position as the target location; Calculate the deviation rate of each monitoring parameter in the deformation feature data and the target deformation feature template data , wherein the deviation rate The mathematical expression is: Where, The first The value of the monitoring parameter, The target deformation feature template data The value of the monitoring parameter, for and The maximum value in ; Deviation rate at each position When both are smaller than a preset deviation rate threshold, it is determined that the deformation feature data is similar to the target deformation feature template data; otherwise, it is determined that the deformation feature data is not similar to the target deformation feature template data.

8. The deformation monitoring and early warning method for an existing tunnel according to claim 1 is characterized in that: Deformation warning for existing tunnels based on the comparison results, including: When the comparison result is similar, the monitoring data sequence is marked as having deformation risk and a warning message is generated, and the warning message is sent to the target object; when the comparison result is dissimilar, the monitoring data sequence is marked as not having deformation risk.

9. The deformation monitoring and early warning method for an existing tunnel according to claim 1, characterized in that: The target location is a pre-marked location where strain is likely to occur.

10. A deformation monitoring and early warning system for an existing tunnel, characterized in that: include: a sample acquisition module, configured to acquire a historical monitoring image sequence of multiple image monitoring points of a reference tunnel, a historical monitoring data sequence of multiple mechanical monitoring locations, and geological structure information of the multiple mechanical monitoring locations, wherein the historical monitoring image sequence includes monitoring images at multiple historical time points, and the historical monitoring data sequence includes the load ratio, temperature and humidity, surrounding rock moisture content, vibration amplitude, and vibration frequency of the lining at multiple historical time points; a deformation event screening module for determining a target location where lining deformation occurs in an existing tunnel and a target time point at which lining deformation occurs based on the historical monitoring image sequence; a template feature extraction module for classifying a plurality of mechanical monitoring locations based on the geological structure information to obtain geological structure categories; and constructing deformation feature template data based on target monitoring data of the mechanical monitoring locations of each geological structure category, wherein the target monitoring data includes monitoring data at a target time point and at multiple historical time points of a target duration before the target time point; A feature extraction module is used to obtain a monitoring data sequence and a geological structure category of a target position of an existing tunnel at multiple time points within a target time period before a current time point, and to extract deformation feature data from the monitoring data sequence; The comparison and warning module is used to determine the target deformation feature template data of the target location based on the geological structure category, perform similarity comparison between the deformation feature data and the target deformation feature template data, and perform deformation warning for the existing tunnel based on the comparison result.

Images