Highway tunnel health monitoring abnormity identification method and system based on neighborhood estimation method

Abstract

The invention relates to a highway tunnel health monitoring abnormity identification method and system based on an estimation neighborhood method. The method comprises the following steps: acquiring monitoring data; performing abnormal value identification on the monitoring data according to a Pauta criterion principle, and recording and storing the abnormal value as first abnormal data; carrying out abnormal value identification on the monitoring data by adopting an LOF algorithm based on density analysis, and recording and storing as second abnormal data; superposing the first abnormal data and the second abnormal data to form third abnormal data; and performing suspicious value verification on the third abnormal data based on an estimation neighborhood method, and taking the abnormal data passing the suspicious value verification as a final identified abnormal value. Compared with the prior art, the method has the advantages of accurate identification result, high reliability and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of highway tunnel monitoring, and in particular to a highway tunnel health monitoring anomaly recognition method and system based on an estimated neighborhood method. Background Art

[0002] When statistically analyzing highway tunnel structure monitoring data, some abnormal data often appear, making it impossible to accurately judge and predict the safety status of the tunnel structure. Highway tunnel structure monitoring abnormal data can usually be divided into gross errors and systematic errors. Gross errors are few in number but large in absolute value. Systematic errors are generally fixed errors generated by instruments and equipment and have certain regularity.

[0003] At present, the commonly used methods for identifying abnormal data include the Raida criterion, the Grubbs criterion, and the t-test method. These methods usually have requirements for the distribution of monitoring data, and statistical methods are not suitable for the identification of multidimensional data. The key technology of the outlier identification method based on statistical analysis is to set a fixed threshold. The Grubbs criterion, the Chauville criterion, and the Dixon criterion need to look up a table based on the number of measurements when setting the threshold. The abnormal data identification results are greatly affected by the number of measurements. The threshold of the Raida criterion does not need to be obtained by looking up a table, but is set based on the mean and variance of the statistical data. However, the Laida criterion is only applicable to the outlier identification of one-dimensional data, while tunnel structure health monitoring data has many influencing factors. When analyzing data, it is usually necessary to consider information such as time, temperature and monitoring location. The Laida criterion cannot be applied to the outlier detection of multidimensional data. The detection models constructed based on distance, density and other methods have different recognition results, resulting in poor accuracy of abnormal data recognition results.

[0004] There are two existing methods for identifying abnormal data in tunnel and bridge structure safety monitoring. One is a traditional single abnormal data identification method based on statistical analysis, and the other is a method that combines multiple types. For example, patent application CN117609907A discloses a method for identifying abnormal values ​​in dam safety monitoring data. By constructing a mutation point detection SNHT model and a local anomaly factor LOF model, abnormal values ​​are identified in safety monitoring data, and the abnormal data identified by the two models are compared, and the common part is extracted as abnormal data. Although the combination of multiple types of abnormal data identification methods can effectively improve the abnormal data identification results, there are differences in the detection results of different identification methods. The existing methods usually take the common part as abnormal data, which easily leads to poor accuracy of abnormal data identification results. In addition, the existing technical solutions lack the step of verifying the detection effect of the monitoring data after abnormal value identification, and the reliability cannot be guaranteed. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a highway tunnel health monitoring anomaly identification method and equipment based on the estimated neighborhood method with accurate identification results and high reliability.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A highway tunnel health monitoring anomaly recognition method based on an estimated neighborhood method comprises the following steps:

[0008] Obtain monitoring data;

[0009] According to the Laida criterion principle, the monitoring data is identified as abnormal values, and recorded and stored as first abnormal data;

[0010] Using a LOF algorithm based on density analysis to identify abnormal values ​​of the monitoring data, and recording and storing the abnormal values ​​as second abnormal data;

[0011] superimposing the first abnormal data and the second abnormal data to form third abnormal data;

[0012] The third abnormal data is subjected to suspicious value verification based on the estimated neighborhood method, and the abnormal data that passes the suspicious value verification is used as the final identified abnormal value.

[0013] Furthermore, the monitoring data is classified and stored according to monitoring section, monitoring item, measuring point location and monitoring time.

[0014] Furthermore, the monitoring items include multiple ones of water and soil pressure outside the lining, lining surface stress, contact pressure, bolt axial force, and joint opening.

[0015] Further, the performing suspicious value check on the third abnormal data based on the estimated neighborhood method includes:

[0016] On the monitoring section, a variogram is established for any two points, and the neighborhood size is determined according to the range value in the variogram at the distance h between the two points.

[0017] The average variance of the observations of all two monitoring points with a distance h in the neighborhood of each outlier is calculated in turn, and then the statistic is obtained;

[0018] It is determined whether the statistic of each abnormal value is greater than a set threshold. If so, the corresponding abnormal value passes the suspicious check.

[0019] Furthermore, the calculation formula of the statistic is:

[0020]

[0021] Wherein, G is any abnormal value in the third abnormal data, I is the statistic of the abnormal value G, M is the arithmetic mean of other observations in the neighborhood that does not contain G, n is the number of observations in the neighborhood that does not contain G, σ 2 is the mean variance.

[0022] Furthermore, the method further comprises:

[0023] The outliers finally identified are removed and filled, and the grey correlation method is used to select the reference sequence, calculate and compare the correlation coefficient between the sequence where the outliers are located and the reference sequence, and the correlation coefficient between the filled data sequence and the reference sequence, to determine the effect of outlier identification.

[0024] Furthermore, the filling is specifically: filling with the mean of the monitoring data before and after the abnormal value.

[0025] Furthermore, the filling is specifically performed by filling through an established interpolation function.

[0026] Furthermore, the selection of the reference sequence is specifically as follows:

[0027] According to the monitoring characteristics of the highway tunnel cross section, a monitoring point is selected on the monitoring section where abnormal data exists, and the monitoring data of the monitoring point with respect to time is used as the reference sequence.

[0028] The present invention also provides a highway tunnel health monitoring anomaly identification system based on the estimated neighborhood method, comprising one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the highway tunnel health monitoring anomaly identification method as described above.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. According to the characteristics of highway tunnel health monitoring, such as multiple monitoring project types, multiple monitoring point locations, and a large amount of monitoring data, the present invention constructs a detection model by integrating the Laida criterion and the LOF algorithm to identify outliers in highway tunnel health monitoring data. The detection model has no distribution requirements and dimensional restrictions on monitoring data, and is suitable for multidimensional data. It can be effectively applied to the identification of outliers in highway tunnel health monitoring to avoid the omission of outlier identification results. In addition, considering the mutual influence of monitoring data from different measuring points in the same area, the estimated neighborhood method in geostatistics is introduced to perform suspicious value inspection on all abnormal data from the perspective of spatial correlation, which can further improve the recognition accuracy.

[0031] 2. According to the distribution characteristics of highway tunnel monitoring data, the present invention introduces the estimated neighborhood method in geostatistics, takes the variation function as the main tool, and performs suspicious value inspection on all identified abnormal data based on the spatial correlation of the monitoring data of the abnormal data measuring point and the measuring point within its influence range. This measure improves the accuracy of the abnormal data identification results.

[0032] 3. The present invention makes full use of the characteristics of the change curve of highway tunnel monitoring data, selects a suitable measuring point on the same monitoring section, uses the monitoring data on the measuring point as a reference sequence, and calculates the correlation coefficient between the monitoring data of the remaining measuring points on the section and the reference sequence. The closer the correlation coefficient is to 1, the more reasonable the monitoring data is. This characteristic can provide a reference for the abnormal value identification effect of highway tunnel health monitoring and improve the reliability of abnormal identification results.

[0033] In summary, the anomaly identification method provided by the present invention is formulated based on the characteristics of highway tunnel health monitoring. In the process, the layout of the monitoring points, the curve change characteristics of the monitoring data, etc. are fully considered; at the same time, in view of the problem that the existing technology does not consider the spatial mutual influence of the civil structure monitoring points, the present invention uses the spatial correlation of the monitoring data at different measuring points on the same monitoring section to perform suspicious value inspection on the identified abnormal data, thereby further improving the accuracy of the identification results. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the process of the present invention;

[0035] Figure 2 2 is a diagram showing a structure of a variation function in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0037] Terminology

[0038] Laida criterion: Data greater than μ+3σ or less than μ-3σ are considered abnormal data and are eliminated. After elimination, the deviation and standard deviation of the remaining measurements are recalculated until all deviations are less than 3σ.

[0039] LOF (Local Outlier Factor) algorithm: measures the degree of abnormality of each data point by comparing the density of its neighboring points. The local density of normal observations is similar to that of their neighboring points, while the local density of abnormal data is relatively small.

[0040] Estimated neighborhood method: The identified observation is placed in the context of a spatially continuous mineralized domain for study. This method considers both the observation itself and several observations within the influence range of the observation.

[0041] Variogram: Variogram is a statistic that describes the spatial correlation of random fields and random processes. It is defined as the variance of the difference between two spatial points. In geostatistics, variogram is used to reconstruct and interpolate spatial fields.

[0042] Grey correlation: Study the influence of complex factors that are interrelated and interact with each other, and determine the essential factors that affect things.

[0043] This embodiment provides a highway tunnel health monitoring anomaly identification method based on the estimated neighborhood method, which realizes the identification of tunnel structure abnormal data through the procedures of abnormal data initial screening, secondary inspection and monitoring data preprocessing effect inspection. Figure 1 As shown, the method comprises the following steps:

[0044] S1. Obtain monitoring data.

[0045] In this embodiment, the monitoring data of different monitoring items about time are obtained through the highway tunnel structure monitoring system, and are classified and stored according to the monitoring section, monitoring item, measuring point location and monitoring time. Among them, the monitoring items include water and soil pressure outside the lining, lining surface stress, contact pressure, bolt axial force, joint opening, etc.

[0046] S2. According to the principle of Laida criterion, an abnormal data detection model based on statistical analysis is constructed, and the abnormal data detection model is used to identify abnormal values ​​of the monitoring data, and the data are recorded and stored as the first abnormal data. At the same time, a LOF algorithm detection model is constructed based on the principle of density analysis, and the LOF algorithm detection model is used to identify abnormal values ​​of the monitoring data, and the data are recorded and stored as the second abnormal data.

[0047] In this embodiment, the steps of identifying abnormal data using the Laida criterion are as follows:

[0048] ① Obtain the highway tunnel structure safety monitoring data about time: x 1 , x 2 , …, x n ;

[0049] ② Calculate the average value μ and standard deviation σ of the above monitoring time series data, the formula is as follows:

[0050]

[0051] ③If the monitoring data x at a certain time k , satisfying |x k-μ>3σ, then the monitoring data x at this time k Recorded as outliers;

[0052] ④ After each abnormal data identification, the remaining measurement values ​​should be re-identified for errors until no abnormal values ​​appear.

[0053] In this embodiment, the steps of identifying outliers using the LOF algorithm detection model are as follows:

[0054] ① Obtain the highway tunnel structure safety monitoring data about time: x 1 , x 2 , …, x n ;

[0055] ② Calculate the distance between each data point and other data points and sort them from small to large;

[0056] ③ Find the reach-dist of the monitoring data point P under a given parameter k. The formula is as follows:

[0057] reach-dist k (O,P)=max{d k (O),d(O,P)}

[0058] In the formula, point P is the monitoring data point that needs to be detected for outliers, point O is the point in the k-distance neighborhood of point P, and d k (O) represents the k-distance of point O, and d(O,P) is the Euclidean distance between point P and point O.

[0059] ④ Calculate the local reachability density of point P, the formula is as follows:

[0060]

[0061] ⑤ Calculate the local anomaly factor value of point P, the formula is as follows:

[0062]

[0063] ⑥If LOF(P) is greater than 1, the recorded point P is an outlier point and the value of point P is an outlier value.

[0064] Highway tunnel safety monitoring data has the characteristics of multiple monitoring item types, multiple monitoring point locations, time series monitoring data, and large data volume. Therefore, the traditional single abnormal data identification method cannot fully meet the above characteristics of highway tunnel monitoring data. This embodiment constructs two different types of outlier detection models based on the Laida criterion and the LOF algorithm to perform outlier detection on the monitoring data.

[0065] The Laida criterion determines the screening threshold through the mean μ and variance σ.k , satisfying x k -μ>3σ, then the monitoring data x k The LOF algorithm determines whether a data point is an outlier by studying the density ratio of each data point to the data points in its neighborhood. Calculate the local reachable density lrd of the data point P k (P), and the average value of the local reachability density of all points in its k-neighborhood Σlrd k (O) / |N k (P)|, the ratio of the average local reachability density of all points in the k-neighborhood of point P to the local reachability density of point P is used as the local outlier factor LOF of point P k (P), if LOF(P) is close to 1, it indicates that the density of point P is close to that of its neighborhood points, and it may belong to the same cluster as its neighborhood points; if LOF(P) is smaller than 1, it indicates that point P is more likely to be in a relatively dense area, that is, an inner point; if LOF(P) is greater than 1, it indicates that point P is more likely to be an outlier.

[0066] The two detection models do not require the distribution of monitoring data, and the LOF algorithm is suitable for outlier identification of high-dimensional data, which effectively improves the accuracy of abnormal data identification.

[0067] S3. Superimpose the first abnormal data and the second abnormal data to form third abnormal data, build a suspicious value detection model based on the estimated neighborhood method, and use the suspicious value detection model to perform suspicious value verification on the third abnormal data, and use the abnormal data that passes the suspicious value verification as the final identified abnormal value.

[0068] In this embodiment, the steps of verifying the third abnormal data by using the estimated neighborhood method are as follows:

[0069] ① Establish a variation function for any two points, the formula is as follows:

[0070]

[0071] Where h is the distance between any two points. Plot the variogram with h as the horizontal coordinate and γ*(h) as the vertical coordinate.

[0072] ② Adjust the distance h and determine the size of the neighborhood m according to the range value in the variation function graph. The variation function graph is as follows: Figure 2 shown.

[0073] ③ Calculate the average variance of the observations of two monitoring points k and l with a distance h in the outlier neighborhood. The formula is as follows:

[0074]

[0075] ④ Calculate the statistic I of the suspicious value G being tested:

[0076]

[0077] Where G is the outlier in the third outlier data; M is the arithmetic mean of other monitoring values ​​in the neighborhood that does not contain G; n is the number of monitoring values ​​in the neighborhood that does not contain G; σ 2 is the average variance of the monitored values ​​in the neighborhood.

[0078] ⑤ Determine whether the statistic I is greater than the set threshold. In this embodiment, the threshold is set to 3.84, that is, when I>3.84, it means that the outlier value G detected in the 95% confidence interval is determined to be abnormal data, then it is determined that the outlier value G in the third abnormal data passes the suspicious test and is determined to be abnormal data.

[0079] After the monitoring data is identified as abnormal data by the Laida criterion and the LOF algorithm, there are certain differences in the identification results of the two methods. This embodiment introduces the estimated neighborhood method in geostatistics to further test all the abnormal data identification results. A suspicious value detection model is constructed based on the estimated neighborhood method, and all abnormal values ​​are re-tested, which further improves the accuracy of the abnormal value identification results.

[0080] When testing abnormal data, the estimated neighborhood method not only studies the abnormal data itself, but also studies other monitoring data within the influence range of the abnormal data. The variance function is the main tool for studying the spatial correlation of random fields and random processes. By adjusting the distance between two points in space, the influence range of the abnormal data can be obtained. The estimated neighborhood method effectively avoids the problem of excessive screening of abnormal data results caused by multiple identification methods, and is more scientific and reasonable than the existing technology of extracting the common part as the abnormal value.

[0081] In a preferred embodiment, in order to improve the reliability of the abnormality identification result, the method further includes:

[0082] S4. The finally identified outliers are removed and filled, and the grey correlation method is used to select a suitable reference sequence, and the correlation coefficient between the sequence where the outliers are located and the reference sequence and the correlation coefficient between the filled data sequence and the reference sequence are calculated and compared to determine the effect of outlier identification.

[0083] In this embodiment, step S4 is specifically as follows:

[0084] ① For all abnormal data x u Remove the outliers and use the mean x of the monitoring data before and after the outliers v to fill in;

[0085] ② According to the monitoring characteristics of highway tunnel cross sections, select appropriate monitoring points on the monitoring sections with abnormal data, and use the monitoring data of the monitoring points in time as the reference sequence;

[0086] ③ Calculate the abnormal data x separately u And the monitoring data x after mean filling v The correlation coefficient with the reference sequence is calculated as follows:

[0087] i. Averaging: Averaging the monitoring data sequence to obtain the averaged reference sequence {x 0 (t)} and p comparison series with abnormal data {x 1 (t), x 2 (t), …, x p (t)}:

[0088] {x o (t)}={x 0 (t 1 ),x 0 (t 2 ),...,x 0 (t n )}

[0089]

[0090] ii. Calculate the correlation coefficient: The correlation coefficient can quantitatively reflect the closeness of two series at a certain moment. Calculate the correlation coefficient ξ at the monitoring time t when abnormal data appears u (t) and ξ v (t), the formula is as follows:

[0091]

[0092] Let the absolute value of the difference between the kth comparison sequence and the reference sequence at time t be Δ(t) = |x 0 (t)-x k (t)|, then ξ u (t) is the monitored abnormal value x in the comparison series at time t u The correlation coefficient with the monitored values ​​in the reference series, ξ v (t) is the mean filling value x in the comparison series at time t v The correlation coefficient with the monitored value in the reference series.

[0093] iii. Compare ξ u (t) and ξ v (t), the comparison result can be used as a reference for the abnormal data identification effect. u (t)<ξ v(t), it indicates that the abnormal data identification result of the monitoring point is relatively accurate. The closer the correlation coefficient is to 1, the more similar the curve shape is, and the better the abnormal value identification effect of the monitoring data is.

[0094] Finally, all the correlation coefficient calculation results are output to provide a reference for the identification of outliers in the monitoring data.

[0095] Characteristics of highway tunnel monitoring data: For the same monitoring section and the same monitoring item, the curves of the monitoring data collected at different monitoring points over time have basically the same shape. Based on this characteristic, this embodiment selects a suitable monitoring point, uses the monitoring data of the monitoring point as a reference sequence, and calculates the correlation coefficient between the monitoring data sequence of the remaining monitoring points and the reference sequence at each monitoring time. The closer the correlation coefficient is to 1, the more normal the monitoring data is. At the same time, the abnormal data is mean-replaced, and the correlation coefficients before and after the replacement are compared. The results can be used as a reference for the abnormal data identification effect. This embodiment verifies the abnormal value identification effect of the monitoring data by calculating the correlation coefficient, thereby improving the reliability of the abnormal identification results.

[0096] In other embodiments, in addition to using the mean value for filling after removing outliers, an interpolation function can also be used for filling. However, establishing a suitable interpolation function is more troublesome than mean filling, and the stable monitoring data curve has small fluctuations, so filling with the mean is simple and reasonable.

[0097] If the above method is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0098] In another embodiment, a highway tunnel health monitoring anomaly identification system based on the estimated neighborhood method is also provided, comprising one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the highway tunnel health monitoring anomaly identification method as described above.

[0099] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0100] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0101] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A highway tunnel health monitoring anomaly identification method based on estimated neighborhood method, characterized in that: The following steps are involved: Obtain monitoring data; According to the Laida criterion principle, the monitoring data is identified as abnormal values, and recorded and stored as first abnormal data; Using a LOF algorithm based on density analysis to identify abnormal values ​​of the monitoring data, and recording and storing the abnormal values ​​as second abnormal data; superimposing the first abnormal data and the second abnormal data to form third abnormal data; The third abnormal data is subjected to suspicious value verification based on the estimated neighborhood method, and the abnormal data that passes the suspicious value verification is used as the final identified abnormal value.

2. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 1 is characterized in that: The monitoring data is classified and stored according to monitoring section, monitoring item, measuring point location and monitoring time.

3. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 1, characterized in that: The monitoring items include multiple ones of the water and soil pressure outside the lining, the surface stress of the lining, the contact pressure, the axial force of the bolts, and the opening of the joints.

4. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 1, characterized in that: The performing suspicious value checking on the third abnormal data based on the estimated neighborhood method comprises: On the monitoring section, a variogram is established for any two points, and the neighborhood size is determined according to the range value in the variogram at the distance h between the two points. The average variance of the observations of all two monitoring points with a distance h in the neighborhood of each outlier is calculated in turn, and then the statistic is obtained; It is determined whether the statistic of each abnormal value is greater than a set threshold. If so, the corresponding abnormal value passes the suspicious check.

5. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 4 is characterized in that: The calculation formula of the statistic is: Wherein, G is any abnormal value in the third abnormal data, I is the statistic of the abnormal value G, M is the arithmetic mean of other observations in the neighborhood that does not contain G, n is the number of observations in the neighborhood that does not contain G, σ 2 is the mean variance.

6. The highway tunnel health monitoring anomaly identification method based on estimated neighborhood method according to claim 1 is characterized in that: The method further includes: The outliers finally identified are removed and filled, and the grey correlation method is used to select the reference sequence, calculate and compare the correlation coefficient between the sequence where the outliers are located and the reference sequence, and the correlation coefficient between the filled data sequence and the reference sequence, to determine the effect of outlier identification.

7. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 6, characterized in that: The filling is specifically: filling with the mean of the monitoring data before and after the abnormal value.

8. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 6, characterized in that: The filling is specifically performed by using an established interpolation function.

9. The method for identifying abnormalities in highway tunnel health monitoring based on the estimated neighborhood method according to claim 6, characterized in that: The selection of the reference sequence is specifically as follows: According to the monitoring characteristics of the highway tunnel cross section, a monitoring point is selected on the monitoring section where abnormal data exists, and the monitoring data of the monitoring point with respect to time is used as the reference sequence.

10. A highway tunnel health monitoring anomaly recognition system based on estimated neighborhood method, characterized in that: The system comprises one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the abnormality identification method for health monitoring of a highway tunnel as claimed in any one of claims 1 to 9.

Images