Engineering monitoring data correction method based on internet of things data platform

Abstract

The present application relates to the technical field of data correction, and particularly relates to an engineering monitoring data correction method based on an Internet of Things data platform, which comprises the following steps: determining the number of monitoring points in each local area according to the obtained point cloud of a building foundation pit in a current period and a historical period; obtaining monitoring data sequences of several dimensions corresponding to the monitoring points of the number of monitoring points, and then determining the abnormal index of each dimension at each time; determining the abnormal state index of each monitoring point according to the abnormal index and the monitoring data, and then screening out several real abnormal time points based on the abnormal state index; determining the soil body damage index of each real abnormal area at the current time according to the abnormal state index and the area of each monitoring point in each real abnormal area of each real abnormal time point; and correcting the monitoring data according to the soil body damage index, which effectively improves the correction effect of the engineering monitoring data and obtains monitoring data with higher numerical accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data correction, in particular to an engineering monitoring data correction method based on an Internet of Things data platform. BACKGROUND

[0002] When monitoring the construction process, the monitoring data is usually analyzed based on an Internet of Things data platform. For example, when monitoring a building foundation pit project, monitoring points are uniformly arranged in the construction range and corresponding sensors are installed to obtain real-time monitoring data of each monitoring point. Subsequently, based on the collected monitoring data, the Internet of Things technology is used to analyze and identify sudden abnormal situations in the construction process of the building foundation pit, and timely output of danger warning information.

[0003] In the actual process of foundation pit construction monitoring, the soil body changes are usually non-uniform. If the above monitoring method is used, the monitoring point setting density may not be suitable for the actual construction situation, and thus the monitoring change trend of the local dangerous area cannot be accurately captured. In the actual construction process, there is a coupling of local soil body rebound caused by soil body load change (such as excavation unloading) and consolidation settlement caused by dewatering, which masks the real sudden abnormality in the construction process.

[0004] The existing global unified algorithm (such as overall Kalman filtering) is used to correct all monitoring data, which cannot distinguish local deformation mechanism and still cannot overcome the defects caused by the fixed monitoring point setting density and the actual local construction situation, resulting in low correction effect of engineering monitoring data. SUMMARY

[0005] In order to solve the technical problems of low correction effect of existing engineering monitoring data and inability to accurately reflect the actual construction situation, the purpose of the present application is to provide an engineering monitoring data correction method based on an Internet of Things data platform. The technical solution adopted is as follows:

[0006] One embodiment of the present application provides an engineering monitoring data correction method based on an Internet of Things data platform, which comprises the following steps:

[0007] Obtain the building foundation pit point cloud of the current period and the historical period when performing inspection on the current building foundation pit construction; the historical period is the previous period of the current period, and the building foundation pit point clouds of the current period and the historical period overlap;

[0008] According to the building foundation pit point clouds of the current period and the historical period, analyze the physical structure complexity of the local area, and determine the number of monitoring points in each local area in the building foundation pit point cloud of the current period;

[0009] acquire a plurality of dimensional monitoring data sequences corresponding to the monitoring points in each local region, and determine an abnormality index of each dimension at each time point according to the monitoring data sequence of each dimension corresponding to each monitoring point;

[0010] determine an abnormality state index of each monitoring point at each to-be-determined abnormal time point according to the abnormality index of each dimension at each time point and the monitoring data, and further screen out a plurality of real abnormal time points based on the abnormality state index;

[0011] acquire each real abnormal region at each real abnormal time point, and determine a soil body damage index of each real abnormal region at the current time point according to the abnormality state index and the area of each monitoring point in each real abnormal region at each real abnormal time point;

[0012] correct the monitoring data of each dimension corresponding to each monitoring point in each real abnormal region at the current time point according to the soil body damage index, to obtain a corrected monitoring data set at the current time point.

[0013] Further, the determination of the monitoring point distribution quantity of each local region in the building foundation pit point cloud of the current period comprises:

[0014] partition the building foundation pit point cloud of the current period and the historical period to obtain each local region corresponding to the current period and the historical period;

[0015] determine the matching region of each local region of the current period in the historical period, and further determine the vector module length of each point in each local region of the current period pointing to the matching point in the corresponding matching region;

[0016] determine the monitoring point distribution index of each local region of the current period according to the area of each local region of the current period and its matching region, the normal vector of each point, and the vector module length of each point in each local region;

[0017] adjust the initial monitoring point distribution quantity of each local region of the current period by using the monitoring point distribution index, to obtain the monitoring point distribution quantity of each local region in the building foundation pit point cloud of the current period.

[0018] Further, the determination of the monitoring point distribution index of each local region of the current period comprises:

[0019] respectively for each local region of the current period, determine a first monitoring point distribution factor of the local region according to the area difference between the local region and its matching region;

[0020] determine a second monitoring point distribution factor of the local region according to the vector module length of each point in the local region;

[0021] determining a third monitoring point distribution factor of the local region according to a normal vector similarity between each point in the local region and its matching point;

[0022] fusing the first monitoring point distribution factor, the second monitoring point distribution factor and the third monitoring point distribution factor of the local region to obtain a monitoring point distribution index of the local region.

[0023] Further, after obtaining the monitoring point distribution quantity of each local region in the building foundation point cloud of the current period, the method further comprises:

[0024] For each local region of the current period, if the monitoring point distribution quantity of the local region is less than the monitoring point distribution quantity of the matching region, the monitoring point distribution of the overlapping region corresponding to the local region and the matching region is kept unchanged, and the monitoring points in the remaining region of the local region except the overlapping region are removed.

[0025] If the monitoring point distribution quantity of the local region is greater than the monitoring point distribution quantity of the matching region, the monitoring point distribution of the overlapping region corresponding to the local region and the matching region is kept unchanged, and the monitoring points in the remaining region of the local region except the overlapping region are added.

[0026] Further, the determination of the abnormality index of each dimension at each time comprises:

[0027] For each monitoring point, a first-order difference sequence of each dimension is determined according to the monitoring data sequence of each dimension.

[0028] A monitoring fluctuation representation value of each dimension at each time is determined according to the first-order difference sequence of each dimension; the monitoring fluctuation representation value is at least used to represent the monitoring fluctuation degree of any time relative to the historical time;

[0029] A monitoring difference representation value of each dimension at each time is determined according to the monitoring data sequence of each dimension; the monitoring difference representation value is at least used to represent the monitoring data difference degree of any time relative to the historical time;

[0030] A monitoring confusion representation value of each dimension at each time is determined according to the fluctuation period corresponding to adjacent extreme value points on the time sequence curve of the monitoring data sequence of each dimension.

[0031] The monitoring fluctuation representation value, the monitoring difference representation value and the monitoring confusion representation value of the same dimension at the same time are fused to obtain the abnormality index of each dimension at each time.

[0032] Further, the determination of the monitoring confusion representation value of each dimension at each time comprises:

[0033] For each dimension, a time period between adjacent extreme points on the time curve is obtained, denoted as a fluctuation period;

[0034] An average value of all fluctuation periods is calculated as a reference fluctuation period;

[0035] A difference between a fluctuation period at each time point and the reference fluctuation period is calculated, denoted as a difference period;

[0036] According to a proportion of the difference period at each time point on the reference fluctuation period, a monitoring confusion representation value of each time point is determined.

[0037] Further, the determination of the abnormal state index of each monitoring point at each pending abnormal time point comprises:

[0038] For each monitoring point, an abnormal threshold is set, a time point with an abnormal index greater than the abnormal threshold is regarded as a pending abnormal time point, and a time point other than the pending abnormal time point is regarded as a normal time point, to obtain each pending abnormal time period and each normal time period of each dimension;

[0039] According to the monitoring data subsequence of each dimension in each pending abnormal time period and each normal time period, a comparison monitoring data subsequence of each pending abnormal time period in each comparison dimension and a comparison monitoring data subsequence of each normal time period in each comparison dimension are determined, the target dimension being any dimension, and the comparison dimension being the remaining dimension other than the target dimension;

[0040] According to the similarity between the monitoring data subsequence of the target dimension in each pending abnormal time period and each comparison monitoring data subsequence of the corresponding pending abnormal time period, a first monitoring similarity index of the target dimension in each pending abnormal time period is determined; similarly, a second monitoring similarity index of the target dimension in the normal time period is determined;

[0041] According to each first monitoring similarity index and second monitoring similarity index of each dimension, an abnormal index sequence in all pending abnormal time periods, and a maximum abnormal dimension number at each pending abnormal time point, an abnormal state index of a monitoring point at each pending abnormal time point is determined.

[0042] Further, the determination of the abnormal state index of the monitoring point at each pending abnormal time point comprises:

[0043] According to the difference between each first monitoring similarity index and second monitoring similarity index of the same dimension, a first abnormal state factor of the monitoring point at each pending abnormal time point is determined;

[0044] obtaining a first-order difference sequence of the abnormal index sequence of each dimension at all pending abnormal periods, and recording a recovery stable time point if at least a preset number of first-order difference values in the first-order difference sequence are not greater than zero continuously;

[0045] determining a second abnormal state factor of the monitoring point at each pending abnormal time point according to a first time length corresponding to the first pending abnormal time point to the recovery stable time point and a second time length corresponding to the current pending abnormal time point of each dimension;

[0046] determining a third abnormal state factor of the monitoring point at each pending abnormal time point according to a comparison between the abnormal index of each dimension at each pending abnormal time point and the abnormal index of the first pending abnormal time point of the corresponding dimension;

[0047] for each pending abnormal time point, obtaining a dimension number corresponding to the pending abnormal time point and each pending abnormal time point before the pending abnormal time point, and taking the maximum dimension number as the maximum abnormal dimension number of the pending abnormal time point;

[0048] fusing the first abnormal state factor, the second abnormal state factor, the third abnormal state factor and the maximum abnormal dimension number of the same pending abnormal time point to obtain an abnormal state index of the monitoring point at each pending abnormal time point.

[0049] Further, the determination of the soil damage index of each real abnormal region at the current time point comprises:

[0050] if the current time point is a real abnormal time point, for each real abnormal region, an abnormal state index sequence of the same monitoring point in the real abnormal region up to the current time point is obtained, and then a first-order difference sequence of the abnormal state index sequence is determined;

[0051] determining an abnormal growth index of each monitoring point in the real abnormal region at the current time point according to the first-order difference sequence of each monitoring point in the real abnormal region at the current time point; the real abnormal region at the current time point is the real abnormal region at the closest abnormal real time point to the current time point;

[0052] matching the real abnormal region at the current time point with each real abnormal region at the first real abnormal time point to obtain a matching abnormal region of the real abnormal region at the current time point;

[0053] determining the soil damage index of the real abnormal region according to the abnormal growth index of each monitoring point in the real abnormal region, the range value corresponding to the abnormal state index of all monitoring points, the area of the real abnormal region and the matching abnormal region.

[0054] Further, the determination of the soil damage index of the real abnormal region comprises:

[0055] Calculate the average value of the abnormal growth index of all monitoring points in the real abnormal area as the first soil mass damage factor of the real abnormal area;

[0056] Corresponding to the range value of the abnormal state index of all monitoring points in the real abnormal area, the second soil mass damage factor of the real abnormal area is obtained;

[0057] The ratio of the area of the real abnormal area to the area of the matching abnormal area is taken as the third soil mass damage factor of the real abnormal area;

[0058] The first soil mass damage factor, the second soil mass damage factor and the third soil mass damage factor of the real abnormal area are fused to obtain the soil mass damage index of the real abnormal area.

[0059] The present application has the following beneficial effects:

[0060] The present application provides an engineering monitoring data correction method based on an Internet of Things data platform, which sets monitoring points according to the soil structure complexity of a real-time three-dimensional model in the foundation pit construction range, and adaptively obtains the monitoring point distribution quantity of different local areas. It can effectively overcome the situation that the fixed monitoring point setting density is not suitable for actual construction, and is beneficial to accurately capturing the monitoring change trend of the local dangerous area. The abnormal state index is determined according to the monitoring data sequence of several dimensions corresponding to each monitoring point, so as to partition the monitoring points at the real abnormal moment and obtain each real abnormal area. It effectively avoids the influence of temporary abnormalities such as soil rebound and consolidation settlement, and is beneficial to accurately identifying real sudden abnormalities. The soil mass damage index is determined according to the abnormal state characteristics and area size of each real abnormal area at each real abnormal moment, so as to correct the monitoring data in real time, effectively improve the correction effect of the monitoring data, and obtain more accurate monitoring data, thereby enhancing the reliability of engineering monitoring and early warning. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0062] Figure 1 A flow chart of an engineering monitoring data correction method based on an Internet of Things data platform is provided for an embodiment of the present application;

[0063] Figure 2 The implementation flow chart of step S2 in the embodiment of the present application;

[0064] Figure 3 An implementation flowchart for determining the abnormal index of each dimension at each time point in an embodiment of the present application is shown in the following table:

[0065] Figure 4 An implementation flowchart for determining the abnormal state index of each monitoring point at each to-be-determined abnormal time point in an embodiment of the present application is shown in the following table:

[0066] Figure 5 An implementation flowchart for determining the soil mass failure index of each real abnormal region at the current time point in an embodiment of the present application is shown in the following table. DETAILED DESCRIPTION

[0067] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined object, the specific implementation, structure, features and effects of the technical solutions proposed by the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0069] The application scenarios to which the present application is directed can be:

[0070] The Internet of Things platform distinguishes different abnormal reasons in the construction process based on the actual monitoring data of the building foundation pit, so as to correct the monitoring data of different local regions, obtain monitoring data that can truly reflect the construction situation, and reduce the possibility of missed or mistaken judgments due to temporary false abnormal phenomena in the construction process, thereby improving the effectiveness of the Internet of Things for building foundation pit monitoring.

[0071] The main purpose of the present application can be:

[0072] The Internet of Things platform accurately corrects the monitoring data in the construction process of the building foundation pit. The monitoring point setting is adjusted according to the actual structural changes, so as to capture the local small deformation and other fluctuations within the construction range; the real soil mass sudden abnormality is determined based on the abnormal fluctuations of the indicators of the adjusted monitoring points; the data of the local region is corrected according to the deterioration trend of the real soil mass sudden abnormality, so that the corrected monitoring data can accurately reflect the actual construction situation.

[0073] One embodiment of the present application provides an engineering monitoring data correction method based on an Internet of Things data platform, as shown in Figure 1 The method comprises the following steps:

[0074] S1, obtain the building foundation point cloud of the current period and the historical period when performing inspection on the current building foundation construction.

[0075] Here, the historical period is the previous period of the current period, the building foundation point cloud is essentially a three-dimensional model, and the building foundation point clouds of the current period and the historical period overlap.

[0076] In one embodiment, the construction scene of the current building foundation is inspected by using an unmanned aerial LiDAR (Laser Imaging Detection and Ranging), i.e., the current building foundation construction scene is monitored once every period, 24 hours is taken as one period, the soil range in the inspection result is marked based on semantic recognition, the remaining points do not participate in subsequent calculation and analysis, and the building foundation point cloud of the current period and the historical period is obtained.

[0077] Further, the initial three-dimensional point cloud captured by the LiDAR at each time is voxel down-sampled, and the ICP (Iterative Closest Point) algorithm is used to splice the point clouds of adjacent time after voxel down-sampling, so as to obtain the complete point cloud of the building foundation in the current period, which is recorded as the building foundation point cloud of the current period. Wherein, the LiDAR is calibrated before inspection. Secondly, it is ensured that the images obtained by the unmanned aerial vehicle at adjacent times have at least 30% overlapping area, and the inspection period is the same within each period, for example, inspection is performed between 14:00 and 15:00 every day. The process of obtaining the building foundation point cloud is prior art, which will not be described here.

[0078] After obtaining the building foundation point cloud of each period and the historical period in real time, it can be transmitted to the Internet of Things data platform for storage and analysis by using wireless transmission technology, so as to realize engineering monitoring.

[0079] It should be noted that the engineering monitoring of the embodiment is to perform inspection while construction, and stop inspection after completion. Due to the analysis and calculation of the distribution quantity of the monitoring points, at least two periods of building foundation point clouds are required, so for the building foundation point cloud of the first day of the construction, i.e., the building foundation point cloud of the first period, the distribution of the monitoring points is not adjusted, and the subsequent monitoring data correction analysis can be directly performed to obtain the monitoring result of the first period; and for the building foundation point cloud of the second period, the distribution of the monitoring points of the second period can be adjusted based on the building foundation point cloud of the first period, and then different dimensions of monitoring data are collected based on the adjusted monitoring points, the monitoring data is corrected first, and then real-time early warning analysis is performed based on the corrected data; the real-time construction early warning analysis is continuously executed until the construction is completed.

[0080] So far, the embodiment obtains the building foundation point cloud of the current period and the last period of the current period for analyzing the distribution density of the monitoring points.

[0081] S2, according to the physical structure complexity of the local area of the building foundation point cloud of the current period and the historical period, determine the number of monitoring points in each local area of the building foundation point cloud of the current period.

[0082] Here, the number of monitoring points is at least used to represent the distribution density of the monitoring points in the local area.

[0083] First of all, it needs to be pointed out that during the actual construction of the building foundation, due to the differences in the stability of the soil in different areas, that is, the complexity of the physical structure characteristics, the degree of engineering monitoring required in different areas is different, that is, for the area with complex structure (for example, the excavation surface has many corners) or the area with soil prone to deformation, the monitoring point density should be increased to facilitate the capture of local small changes. Therefore, it is necessary to adjust the monitoring point distribution density in different local areas of the building foundation point cloud of the current period based on the physical structure characteristics of the building foundation of the adjacent two periods.

[0084] It needs to be further pointed out that, first, the point cloud density is used to divide the three-dimensional model of the current period and the historical period, if the physical structure of a local area in the building foundation point cloud of the current period is relatively complex, and there is a large deformation relative to the local area of the historical period, the monitoring point of the local area of the current period is increased.

[0085] As an exemplary embodiment, the above step S2 can be realized by the steps shown in the following table: Figure 2

[0086] S21, divide the building foundation point cloud of the current period and the historical period to obtain each local area corresponding to the current period and the historical period.

[0087] Due to the differences in the stability of the soil in different areas and the complexity of the physical structure characteristics, if the monitoring points are uniformly set for the entire building foundation point cloud, it may lead to that the distribution density of the monitoring points cannot be suitable for the real-time construction situation, which is not conducive to capturing the monitoring change trend of the local dangerous area. Here, the building foundation point cloud is first divided into zones.

[0088] In one embodiment, DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is used to divide the building foundation point cloud of the current period and the historical period, respectively, to obtain each local area in the building foundation point cloud of the current period and each local area in the building foundation point cloud of the historical period.​

[0089] wherein the neighborhood radius is set as 1.5 meters, which can be adjusted according to the actual foundation pit range, and the minimum point number of each local area is set as 50. The implementation process of DBSCAN clustering is prior art, which is not described here.

[0090] S22, determining the matching area of each local area of the current period in the history period, and further determining the vector module length of each point in each local area of the current period pointing to the matching point in the corresponding matching area.

[0091] First step, determining the matching area of each local area of the current period in the history period.

[0092] In order to analyze the deformation degree and structural complexity of similar local areas in adjacent periods, it is necessary to register each local area of the current period and the history period.

[0093] In one embodiment, for each local area respectively, the local area of the current period is matched with the local area of the history period with the maximum overlapping area, to obtain the matching area of the local area of the current period in the history period.

[0094] It is worth noting that for the local areas without overlapping, the subsequent monitoring data correction analysis is not performed.

[0095] Second step, determining the vector module length of each point in each local area of the current period pointing to the matching point in the corresponding matching area.

[0096] First sub-step, determining the matching point of each point in each local area of the current period.

[0097] In one embodiment, a point in the local area of the current period is matched with the point in the corresponding matching area with the closest coordinate distance, and the point in the matching area with the closest coordinate distance is taken as the matching point of the point.

[0098] Second sub-step, obtaining the vector module length of each point in each local area of the current period.

[0099] In one embodiment, according to the coordinate positions of each point and its matching point, the vector of each point pointing to the corresponding matching point in each local area of the current period is determined, and then the vector module length of each point is calculated.

[0100] S23, determining the monitoring point distribution index of each local area of the current period according to the area of each local area of the current period and its matching area, the normal vector of each point, and the vector module length of each point in each local area.

[0101] Here, the monitoring point distribution index is used to at least represent the degree of engineering monitoring of the local region, and the larger the monitoring point distribution index, the more the local region needs to be monitored, and the greater the monitoring point distribution density of the local region should be.

[0102] As an exemplary embodiment, the monitoring point distribution index of each local region in the current cycle is determined, including:

[0103] First, according to the area difference between the local region and its matching region, the first monitoring point distribution factor of the local region is determined.

[0104] In one embodiment, the absolute value of the difference between the area of the local region and the area of its matching region is calculated, and the absolute value of the difference between the two areas is taken as the first monitoring point distribution factor.

[0105] In another embodiment, the square of the difference between the area of the local region and the area of its matching region is calculated, and the square of the difference between the two areas is taken as the first monitoring point distribution factor.

[0106] Second, according to the vector length of each point in the local region, the second monitoring point distribution factor of the local region is determined.

[0107] In one embodiment, the average value of the vector length of all points in the local region is calculated as the second monitoring point distribution factor.

[0108] In another embodiment, the cumulative sum of the vector length of all points in the local region is calculated as the second monitoring point distribution factor.

[0109] Third, according to the normal vector similarity between each point in the local region and its matching point, the third monitoring point distribution factor of the local region is determined.

[0110] In one embodiment, the average value of the cosine similarity between the normal vector of each point in the local region and the normal vector of its matching point is calculated, and the average value of all cosine similarities is negatively correlated, such as taking the inverse, to obtain a negatively correlated value as the third monitoring point distribution factor of the local region. The calculation process of the cosine similarity is a prior art, which is not described here.

[0111] Fourth, the first monitoring point distribution factor, the second monitoring point distribution factor and the third monitoring point distribution factor of the local region are fused to obtain the monitoring point distribution index of the local region.

[0112] In one embodiment, the calculation formula of the monitoring point distribution index of the jth local region can be:

[0113] ; in the formula, an index of distribution of monitoring points representing the jth local region, a second index of distribution of monitoring points representing the jth local region, an area of a matching region representing the jth local region, an area of the jth local region, an absolute value function, a first index of distribution of monitoring points representing the jth local region, an average value of cosine similarity between normal vectors of each point in the jth local region and normal vectors of matching points thereof, a third index of distribution of monitoring points representing the jth local region.

[0114] In the calculation formula of the index of distribution of monitoring points, each parameter is normalized before calculation to overcome the influence of different dimensions; The greater the value is, the greater the deformation degree of the local region compared with the matching region is; The smaller the value is, the higher the structural complexity of the local region is. Generally, the smaller the value is, the more complex the structure of the local region is. There is no possibility of being zero. If there is an extreme case, a non-zero constant such as 0.001 is added at the position of the denominator of the fraction. With respect to the historical period, if the deformation degree of the local region of the current period is greater and the internal structural complexity is higher, it is indicated that the local region of the current period needs to be monitored more, the index of distribution of monitoring points is greater, and it is further indicated that the density of monitoring points in the local region should be higher.

[0115] Referring to the calculation process of the index of distribution of monitoring points of the jth local region of the current period, the index of distribution of monitoring points of each local region of the current period can be obtained.

[0116] S24, adjusting the initial distribution number of monitoring points of each local region of the current period by using the index of distribution of monitoring points to obtain the distribution number of monitoring points of each local region in the point cloud of the building foundation pit of the current period.

[0117] In an embodiment, the calculation formula of the distribution number of monitoring points of the jth local region can be:

[0118] In the formula, T represents the current period, represents the distribution number of monitoring points of the jth local region of the current period, and T represents the current period, is used to realize normalization processing, and the value range is limited to -1 to 1, represents the index of distribution of monitoring points of the jth local region of the current period, represents a rounding-up function.

[0119] In another embodiment, when T is an integer greater than or equal to 3, the calculation formula of the number of monitoring point distributions of the jth local area can also be:

[0120] wherein, represents the historical period, represents the initial number of monitoring point distributions of the jth local area in the current period, represents the monitoring point distribution index of the jth local area in the historical period, that is, the monitoring point distribution index of the jth local area in the current period in the matching area in the historical period, represents the range of the monitoring point distribution indexes of all matching areas matching the jth local area in the current period in all periods before the current period.

[0121] In the calculation formula of the number of monitoring point distributions, the initial number of monitoring point distributions can be equal to the number of monitoring point distributions of the matching area in the historical period of the jth local area in the current period, and can also be equal to the artificially preset number of monitoring point distributions. and are both monitoring point distribution correction coefficients, used to adjust the initial number of monitoring point distributions of the jth local area in the current period.

[0122] Referring to the calculation process of the number of monitoring point distributions of the jth local area in the current period described above, the number of monitoring point distributions of each local area in the current period can be obtained.

[0123] After obtaining the number of monitoring point distributions of each local area in the building foundation point cloud in the current period, the following steps are further included:

[0124] For each local area in the current period, if the number of monitoring point distributions of the local area is less than the number of monitoring point distributions of the matching area, the number of monitoring points in the overlapping area corresponding to the local area and the matching area is kept unchanged, and the monitoring points in the remaining area corresponding to the local area except the overlapping area are screened out according to the number of monitoring points to be reduced.

[0125] If the number of monitoring point distributions of the local area is greater than the number of monitoring point distributions of the matching area, the number of monitoring points in the overlapping area corresponding to the local area and the matching area is kept unchanged, and the monitoring points in the remaining area corresponding to the local area except the overlapping area are added according to the number of monitoring points to be added.

[0126] It is worth noting that the positions of the screened out or added monitoring points are evenly distributed in the local area.

[0127] At this point, the distribution of each monitoring point in each local area in the current period is obtained, and sensors are installed at each monitoring point to facilitate the subsequent collection of monitoring data of different dimensions.

[0128] S3, obtaining a plurality of dimension monitoring data sequences corresponding to the monitoring points in each local area, and determining an abnormal index of each dimension at each time according to the monitoring data sequence of each dimension corresponding to each monitoring point.

[0129] Here, the abnormal index is used to at least represent the abnormal situation of each dimension monitoring at each time in the preset current period. The larger the abnormal index of a certain dimension at a certain time, the greater the construction interference of the dimension on the monitoring point at that time, and the greater the possibility of the time being an abnormal time.

[0130] First of all, it should be pointed out that in the actual construction process, the monitoring data collected by the sensors corresponding to each monitoring point has certain fluctuations, but the fluctuation amplitude is usually small and can gradually recover to a stable state. However, when there is construction interference, such as excavation unloading or disassembly of supporting structure causing surface uplift in some areas; due to precipitation or pore water discharge causing local soil compression; real fluctuations such as soil instability cause sudden increase of local displacement and stress. Therefore, the abnormal index of each time is determined based on the fluctuation difference of the monitoring data of different dimensions collected by the monitoring points.

[0131] Obtaining a plurality of dimension monitoring data sequences corresponding to the monitoring points in each local area, comprising:

[0132] In one embodiment, the monitoring points of the monitoring point distribution quantity calculated in the above step S2 are uniformly installed in the local area, and different types of sensors are installed at each monitoring point: inclinometer for monitoring horizontal displacement, static leveling instrument for monitoring vertical settlement, strain gauge for monitoring supporting structure stress, vibration sensor for monitoring sudden vibration, and soil moisture content sensor for monitoring monitoring point water content; Through different types of sensors, the monitoring data sequence of each dimension corresponding to each monitoring point can be obtained, and the monitoring data sequence is composed of monitoring data at each time in the preset current period.

[0133] Among them, the monitoring frequency can be set to collect once a minute, and the time stamps of the monitoring data obtained by the sensors corresponding to the same monitoring point are aligned; the current preset period can be set to 1 hour. The parameter size when collecting monitoring data can be set by the implementer according to the specific actual situation, which is not limited here.

[0134] After obtaining the monitoring data sequences of different dimensions, wireless transmission technology can be used to transmit to the Internet of Things data platform for storage and analysis to determine the soil damage index for correcting the monitoring data.

[0135] As an exemplary embodiment, for each monitoring point, the abnormality index of each dimension at each time is determined by Figure 3 as shown in the steps.

[0136] The monitoring data in any dimension is analyzed. The greater the difference between the monitoring data at a certain time and the fluctuation amplitude of the historical monitoring data, the greater the degree of deviation of the monitoring data from the historical monitoring mean (i.e., the historical stable state), and the less consistent the fluctuation duration is with the historical fluctuation rule, the greater the abnormality index at the time, and the greater the possibility of the time being an abnormal time.

[0137] S31, according to the monitoring data sequence of each dimension, a first-order difference sequence of each dimension is determined.

[0138] In an embodiment, for each dimension, the first-order difference values of the monitoring data sequence are obtained based on the monitoring data sequence to form a first-order difference sequence. The calculation process of the first-order difference is a prior art, which is not described here.

[0139] S32, according to the first-order difference sequence of each dimension, a monitoring fluctuation representation value of each dimension at each time is determined.

[0140] Here, the monitoring fluctuation representation value is at least used to represent the monitoring fluctuation degree at any time relative to the historical time.

[0141] In an embodiment, the calculation formula of the monitoring fluctuation representation value of the zth dimension at the ith time can be:

[0142] ; in which, represents the monitoring fluctuation representation value of the zth dimension at the ith time, represents the first-order difference value of the zth dimension at the ith time, represents the average value of the first-order difference values of all historical times before the ith time of the zth dimension, represents the absolute value function.

[0143] S33, according to the monitoring data sequence of each dimension, a monitoring difference representation value of each dimension at each time is determined.

[0144] Here, the monitoring difference representation value is at least used to represent the monitoring data difference degree at any time relative to the historical time.

[0145] In an embodiment, the calculation formula of the monitoring difference representation value of the zth dimension at the ith time can be:

[0146] ; in which, represents the monitoring difference representation value of the zth dimension at the ith time, represents the monitoring data of the zth dimension at the ith time point, represents the average value of the detection data of the zth dimension at all historical time points before the ith time point.

[0147] S34, a time sequence curve of the monitoring data sequence of each dimension is obtained, and according to the fluctuation period corresponding to adjacent extreme value points on the time sequence curve of each dimension, a monitoring chaos representation value of each dimension at each time point is determined.

[0148] In an embodiment, for each dimension, a least square method is used to perform curve fitting on the monitoring data sequence to obtain the time sequence curve of the monitoring data sequence.

[0149] As an exemplary embodiment, for each dimension respectively, determining the monitoring chaos representation value at each time point includes:

[0150] First, the period between adjacent extreme value points on the time sequence curve is obtained, denoted as a fluctuation period.

[0151] In an embodiment, each extreme value point on the time sequence curve is labeled, and the period between adjacent two extreme value points is taken as a fluctuation period.

[0152] Second, the average value of all fluctuation periods is calculated as a reference fluctuation period.

[0153] Third, the difference between the fluctuation period at each time point and the reference fluctuation period is calculated, denoted as a difference period.

[0154] In an embodiment, the absolute value of the difference between the fluctuation period and the reference fluctuation period is calculated as the difference period.

[0155] Fourth, according to the proportion of the difference period at each time point in the reference fluctuation period, the monitoring chaos representation value at each time point is determined.

[0156] Here, the monitoring chaos representation value is used at least to represent the inconsistency between the fluctuation period at each time point and the historical fluctuation rule.

[0157] In an embodiment, for each time point, the ratio of the difference period to the reference fluctuation period is taken as the monitoring chaos representation value.

[0158] S35, the monitoring fluctuation representation value, the monitoring difference representation value and the monitoring chaos representation value of the same time point of the same dimension are fused to obtain an abnormal index of each dimension at each time point.

[0159] Here, the monitoring fluctuation representation value, the monitoring difference representation value and the monitoring chaos representation value all present positive correlation with the abnormal index, and the positive correlation means that the larger the independent variable, the larger the dependent variable.

[0160] In one embodiment, the calculation formula of the anomaly index of the zth dimension at the ith time point can be:

[0161] In the formula, denotes the anomaly index of the zth dimension at the ith time point, norm denotes a linear normalization function, denotes the monitoring fluctuation characteristic value of the zth dimension at the ith time point, denotes the monitoring difference characteristic value of the zth dimension at the ith time point, denotes a reference fluctuation period, denotes the fluctuation period in which the zth dimension is located at the ith time point, denotes the monitoring confusion characteristic value of the zth dimension at the ith time point.

[0162] In the calculation formula of the anomaly index, only the numerical value is taken for analysis when the fusion analysis of each characteristic value is performed, and the influence of the unit is not considered. Moreover, the fusion result of each characteristic value is normalized, which can overcome the influence of the dimension or unit and ensure that the anomaly index is a dimensionless degree value.

[0163] According to the above calculation process of the anomaly index of the zth dimension at the ith time point, the anomaly index of each dimension at each time point can be obtained.

[0164] Thus, the quantification index for representing the possibility of an abnormal time point, i.e., the anomaly index of each dimension at each time point, is obtained in this embodiment.

[0165] S4, according to the anomaly index of each dimension at each time point and the monitoring data, determines the anomaly state index of each monitoring point at each to-be-determined abnormal time point, and then screens out a plurality of real abnormal time points based on the anomaly state index.

[0166] Here, the anomaly state index is at least used to represent the possibility that the monitoring point is in a real abnormal state. The greater the anomaly state index, the greater the possibility that the monitoring point at the to-be-determined abnormal time point is in a real abnormal time point, that is, the greater the possibility that the to-be-determined abnormal time point is a real abnormal time point.

[0167] First of all, it needs to be pointed out that, under normal circumstances, temporary abnormal fluctuations (such as soil rebound and consolidation settlement) only cause abnormal monitoring of a small part of the dimensions, at this time, the ability to maintain the stability of the soil at the corresponding monitoring point position is still strong, the damage degree of the correlation between the dimensions is smaller, and it can gradually recover to a stable state; while the building foundation is in a real abnormal state (such as soil instability, etc.), at this time, the ability of the soil to maintain stability is usually weak, and the correlation between the dimensions cannot continue to be maintained, that is, usually, a large number of dimensions are simultaneously monitored, and the abnormal fluctuations are sudden, the correlation is poor, and cannot be recovered.

[0168] For a single monitoring point, if the abnormality degree of each dimension corresponding to the monitoring point is low and there is no obvious accumulation, and the monitoring point can quickly recover to a stable state, the abnormality of the monitoring point is more likely to be soil rebound; when the abnormality degree has a certain accumulation and the time required to recover to a stable state is long, but the dimensions during the abnormal period can still maintain stability, the abnormality of the monitoring point is more likely to be consolidation settlement; the real abnormality refers to that all dimensions cannot maintain correlation and the abnormality degree has poor stability and rapid increase, and cannot recover to a stable state, at this time, it is more likely that the monitoring point has a real abnormality. Therefore, based on the above analysis, the temporary abnormality and the real abnormality are distinguished, that is, all real abnormal time points are screened out.

[0169] As an exemplary embodiment, for each monitoring point, the abnormality state index of each monitoring point at each pending abnormal time point is determined, and the abnormality state index is determined by Figure 4 The steps shown in the figure are implemented.

[0170] S41, setting an abnormal threshold, taking the time point with an abnormal index greater than the abnormal threshold as a pending abnormal time point, and taking the time point other than the pending abnormal time point as a normal time point, to obtain each pending abnormal time period and each normal time period of each dimension.

[0171] In an embodiment, for each dimension, the abnormal threshold can be set to 0.5, the time point with an abnormal index greater than the abnormal threshold 0.5 is taken as a pending abnormal time point, and the time point other than the pending abnormal time point is taken as a normal time point; then, the continuously adjacent pending abnormal time points are merged into a pending abnormal time period, and the continuously adjacent normal time points are merged into a normal time period, to obtain each pending abnormal time period and each normal time period of each dimension.

[0172] The size of the abnormal threshold can be set by the implementer according to the specific actual situation, which is not limited here.

[0173] S42, according to the monitoring data subsequence of each dimension in each pending abnormal time period and each normal time period, determining the comparison monitoring data subsequence of each pending abnormal time period of the target dimension in each comparison dimension, and the comparison monitoring data subsequence of each normal time period in each comparison dimension.

[0174] Here, the target dimension is any dimension, and the comparison dimension is the remaining dimension except the target dimension.

[0175] In one embodiment, for each pending abnormal period of the target dimension, each comparative dimension obtains a subsequence of monitoring data at the same period as the pending abnormal period of the target dimension, as a comparative subsequence of monitoring data of the pending abnormal period in the comparative dimension; for each normal period of the target dimension, each comparative dimension obtains a subsequence of monitoring data at the same period as the normal period of the target dimension, as a comparative subsequence of monitoring data of the normal period in the comparative dimension.

[0176] For example, a certain normal period of the target dimension is from 1:20 to 1:23, and a subsequence of monitoring data of each comparative dimension at 1:20 to 1:23 is obtained, as a comparative subsequence of monitoring data of the normal period in the comparative dimension.

[0177] S43, according to the similarity between the subsequence of monitoring data of the target dimension at each pending abnormal period and the comparative subsequence of monitoring data of each comparative dimension at the corresponding pending abnormal period, a first monitoring similarity index of the target dimension at each pending abnormal period is determined; similarly, a second monitoring similarity index of the target dimension at the normal period is determined.

[0178] Here, the first monitoring similarity index is used to represent at least the correlation strength between the monitoring situation of the target dimension at the pending abnormal period and the monitoring situation of each comparative dimension at the same period.

[0179] In one embodiment, taking a certain pending abnormal period as an example, a subsequence of monitoring data of the certain pending abnormal period and each comparative subsequence of monitoring data are obtained, and the Pearson correlation coefficients of the subsequence of monitoring data and each comparative subsequence of monitoring data are calculated, and the average of all Pearson correlation coefficients is taken as the first monitoring similarity index of the target dimension at the pending abnormal period. The calculation process of the Pearson correlation coefficient is a prior art, which is not described here.

[0180] In one embodiment, a subsequence of monitoring data of a certain normal period and each comparative subsequence of monitoring data are obtained, and the Pearson correlation coefficients of the subsequence of monitoring data and each comparative subsequence of monitoring data are calculated; the average of all Pearson correlation coefficients is calculated to obtain the average correlation coefficient of the certain normal period; and the average of the average correlation coefficients of all normal periods is taken as the second monitoring similarity index of the target dimension at the normal period.

[0181] Referring to the above obtaining process of each first monitoring similarity index and second monitoring similarity index of the target dimension, each first monitoring similarity index and second monitoring similarity index of each dimension can be obtained.

[0182] S44. Based on the first and second monitoring similarity indicators of each dimension, the sequence of abnormal indicators in all pending abnormal periods, and the maximum number of abnormal dimensions at each pending abnormal time, determine the abnormal status indicators of the monitoring point at each pending abnormal time.

[0183] As an exemplary implementation, determining the abnormal state indicators of the monitoring point at each pending abnormal time includes:

[0184] The first step is to determine the first abnormal state factor of the monitoring point at each pending abnormal time based on the differences between the first and second monitoring similar indicators of the same dimension.

[0185] Here, the first abnormal state factor is used to characterize the degree of stability of the correlation maintained between different dimensions of monitoring. When the first abnormal state factor is small, it indicates that the monitoring situation of each dimension can maintain a relatively stable correlation.

[0186] In one embodiment, for each pending anomaly time, the first monitoring similarity index of each pending anomaly time period before the pending anomaly time is obtained; the absolute value of the difference between each first monitoring similarity index and the second monitoring similarity index before the pending anomaly time under the same dimension is calculated, and the average of all absolute values ​​of the difference is used as the correlation difference index to obtain the correlation difference index of each dimension at the pending anomaly time; the cumulative value of the correlation difference index of all dimensions at the pending anomaly time is calculated as the first anomaly state factor of the monitoring point at the pending anomaly time.

[0187] As an example, the monitoring point is at the... The formula for calculating the first anomalous state factor at each undetermined anomalous time can be:

[0188] In the formula, Indicates the monitoring point at the 1st The first anomalous state factor at each pending anomalous time, where Z represents the number of dimensions and z represents the dimension index. This indicates that the z-th dimension is located at the position of the z-th dimension. The s-th first monitoring similarity indicator before the pending anomaly time. This represents the second monitoring similarity index in the z-th dimension. This indicates that the z-th dimension is in the z-th position. The correlation difference index of each pending abnormal moment.

[0189] In the formula for calculating the first abnormal state factor, The larger the value, the more likely the z-th dimension is to be located at the position of the z-th dimension. The more severe the interference with the correlation between the various undetermined anomaly periods before the first undetermined anomaly time and the comparison dimension, the more severe the interference. each pending abnormal time period containing the pending abnormal time period in which the first The first abnormal state factor of the monitoring point at each pending abnormal time period is the same.

[0190] In a second step, a first-order differential sequence of the abnormal index sequence of each dimension in all pending abnormal time periods is obtained, and if the first-order differential value of at least a preset number of continuous time points in the first-order differential sequence is not greater than zero, the corresponding time point is recorded as a recovery stable time point.

[0191] In an embodiment, the preset number can be set to 3, which is not specifically limited here.

[0192] In a third step, a second abnormal state factor of the monitoring point at each pending abnormal time point is determined according to a first length corresponding to the first pending abnormal time point to the recovery stable time point and a second length corresponding to the first pending abnormal time point to the current pending abnormal time point of each dimension.

[0193] Here, the second abnormal state factor is at least used to represent the efficiency of the recovery stable state after being marked as an abnormal situation; the first length is the length corresponding to the first pending abnormal time point to the recovery stable time point of each dimension, and the second length is the length corresponding to the first pending abnormal time point to the current pending abnormal time point of each dimension; the current pending abnormal time point is any pending abnormal time point.

[0194] In an embodiment, the average value of the first length of all dimensions is calculated for each pending abnormal time point, and the average value of the second length of all dimensions is calculated, and the ratio of the average value of the first length to the average value of the second length is taken as the second abnormal state factor of the monitoring point at the corresponding pending abnormal time point.

[0195] It is worth noting that if a dimension does not exist to the stable state, the length between the first marked pending abnormal time point and the current pending abnormal time point of the dimension is taken as the first length.

[0196] In a fourth step, a third abnormal state factor of the monitoring point at each pending abnormal time point is determined according to the comparison of the abnormal index of each dimension at each pending abnormal time point and the abnormal index of the first pending abnormal time point of the corresponding dimension.

[0197] Here, the third abnormal state factor is at least used to represent the cumulative situation of the abnormal degree up to the pending abnormal time point.

[0198] In one embodiment, for each pending abnormal time, a ratio of the abnormality indicator of each dimension at the pending abnormal time to the abnormality indicator of the corresponding dimension at the first pending abnormal time is determined, and an average of the ratios of all dimensions is taken as the third abnormality state factor of the monitoring point at the pending abnormal time.

[0199] In one embodiment, for each pending abnormal time, the number of dimensions corresponding to the pending abnormal time and each pending abnormal time before the pending abnormal time is obtained, and the maximum number of dimensions is taken as the maximum abnormal dimension number of the pending abnormal time.

[0200] Here, the greater the maximum abnormal dimension number, the more likely the monitoring abnormality up to the pending abnormal time exhibits unstable and rapidly growing data characteristics, and the more likely the pending abnormal time is a real abnormal time.

[0201] In one embodiment, the number of dimensions of the abnormality is obtained at the pending abnormal time and each pending abnormal time before the pending abnormal time, and the maximum number of dimensions is taken as the maximum abnormal dimension number of the pending abnormal time.

[0202] For example, assuming there are 3 dimensions, the first time, the second time, and the fourth time of the first dimension are pending abnormal times, the pending abnormal times corresponding to the first dimension are marked as , , ; the first time, the second time, and the third time of the second dimension are pending abnormal times, the pending abnormal times corresponding to the second dimension are marked as , , ; the first time of the third dimension is a pending abnormal time, and the pending abnormal time corresponding to the third dimension is marked as . At this time, the pending abnormal time corresponds to three dimensions, the pending abnormal time corresponds to two dimensions, the pending abnormal time corresponds to one dimension, and the pending abnormal time corresponds to one dimension, then the maximum abnormal dimension number of the fourth pending abnormal time is 3.

[0203] In one embodiment, for each pending abnormal time, a ratio of the abnormality indicator of each dimension at the pending abnormal time to the abnormality indicator of the corresponding dimension at the first pending abnormal time is determined, and an average of the ratios of all dimensions is taken as the third abnormality state factor of the monitoring point at the pending abnormal time.

[0204] Here, the first abnormality state factor, the second abnormality state factor, the third abnormality state factor, and the maximum abnormal dimension number are positively correlated with the abnormality state indicator.

[0205] In one embodiment, the abnormal state index of the monitoring point at the i-th pending abnormal moment can be calculated according to the following formula:

[0206] ; wherein, the abnormal state index of the monitoring point at the i-th pending abnormal moment is represented by norm represents a linear normalization function for limiting data between 0 and 1, the first abnormal state factor of the monitoring point at the i-th pending abnormal moment is represented by the average value of the first time length of different dimensions at the i-th pending abnormal moment is represented by the average value of the second time length from the first pending abnormal moment to the i-th pending abnormal moment in different dimensions is represented by the second abnormal state factor of the monitoring point at the i-th pending abnormal moment is represented by the third abnormal state factor of the monitoring point at the i-th pending abnormal moment is represented by the maximum abnormal dimension number at the i-th pending abnormal moment is represented by

[0207] It should be noted that in the calculation formula of the abnormal state index, only the numerical values of different calculation parts are fused and analyzed, and the unit influence is not considered for normal multiplication calculation.

[0208] As an exemplary embodiment, a plurality of real abnormal moments are screened based on the abnormal state index, including:

[0209] In one embodiment, for the i-th pending abnormal moment, if , the i-th pending abnormal moment is a real abnormal moment; if , the i-th pending abnormal moment is a consolidation settlement moment; if , the i-th pending abnormal moment is a soil rebound fluctuation moment. wherein, 0.4 and 0.7 are abnormal state threshold values, and the implementer can set the numerical value of the abnormal state threshold value according to the specific actual situation, which is not specifically limited here.

[0210] Thus, the present embodiment obtains each real abnormal moment in the preset current period.

[0211]

[0212] ​​​​​​​​​​​​S5, obtaining each real abnormal region at each real abnormal moment, and determining the soil damage index of each real abnormal region at the current moment according to the abnormal state index and area of each monitoring point in each real abnormal region at each real abnormal moment.

[0213] Firstly, each real abnormal region at each real abnormal moment is obtained.

[0214] In an embodiment, for each real abnormal moment, the monitoring points at the real abnormal moment are taken as abnormal monitoring points, and if all the adjacent monitoring points are abnormal monitoring points, they are merged into one region as a real abnormal region, that is, the region growing algorithm is used to process all the abnormal monitoring points existing at the same real abnormal moment, so that each real abnormal region at each real abnormal moment can be obtained.

[0215] Secondly, the soil damage index of each real abnormal region at the current moment is determined according to the abnormal state index and area of each monitoring point in each real abnormal region at each real abnormal moment, as shown in formula (2). Figure 5

[0216] In actual construction process, when there is a real abnormality locally, it usually has the possibility of damage transmission to the surrounding area, for example, there may be a loss of surrounding soil or stress redistribution in the adjacent area, so that there is a certain expansion in space locally. If, by the current moment, a real abnormal region has a large expansion trend, and the abnormal degree of each monitoring point in the real abnormal region has a large growth and is unstable, the soil damage degree of the real abnormal region is large, and it is more necessary to increase the strength of data correction of the monitoring points in the real abnormal region.

[0217] S51, if the current moment is a real abnormal moment, then for each real abnormal region, the abnormal state index sequence of the same monitoring point in the real abnormal region by the current moment is obtained, and then the first-order difference sequence of the abnormal state index sequence is determined.

[0218] In an embodiment, the abnormal state index of each monitoring point in the real abnormal region at different real abnormal moments before the current moment is counted, and the abnormal state index sequence corresponding to each monitoring point in the real abnormal region is obtained.

[0219] If the current moment is not a real abnormal moment, the soil damage index analysis is not needed, and the inverse distance weighting (IDW) is directly used to correct and process the monitoring data of each monitoring point in different dimensions at the current moment, specifically as follows:

[0220] ​In one embodiment, taking a single monitoring point as an example, taking all adjacent normal monitoring points of the monitoring point as reference monitoring points, respectively acquiring IDW basic weights of the reference monitoring points relative to the monitoring point, and taking the average of all IDW basic weights as the true IDW weight of the monitoring point, the monitoring data of the monitoring point in different dimensions at the current time is dynamically weighted and corrected by IDW. The implementation process of IDW is prior art, which will not be described here.

[0221] S52, determining the abnormal growth index of each monitoring point in the true abnormal area at the current time according to the first-order difference sequence of each monitoring point in the true abnormal area at the current time.

[0222] Here, the greater the abnormal growth index, the greater the growth degree of the abnormal degree of the monitoring point in the true abnormal area over time, and the greater the soil damage degree of the true abnormal area. If the current time is not the true abnormal time, the true abnormal area at the current time is the true abnormal area of the nearest abnormal true time at the current time.

[0223] In one embodiment, the average of all first-order difference values in the first-order difference sequence of the monitoring point is taken as the abnormal growth index of the corresponding monitoring point.

[0224] S53, matching the true abnormal area at the current time with each true abnormal area at the first true abnormal time to obtain the matching abnormal area of the true abnormal area at the current time.

[0225] Here, the first true abnormal time refers to the first true abnormal time in the current preset period.

[0226] In one embodiment, the overlapping range between the true abnormal area at the current time and each true abnormal area at the first true abnormal time is determined, and the true abnormal area of the first true abnormal time with the largest overlapping range is taken as the matching abnormal area of the true abnormal area at the current time.

[0227] S54, determining the soil damage index of the true abnormal area according to the abnormal growth index of each monitoring point in the true abnormal area, the range value corresponding to the abnormal state index of all monitoring points, the area of the true abnormal area and the matching abnormal area.

[0228] Here, the soil damage index is at least used to represent the degree of soil damage in the true abnormal area. The greater the soil damage index, the more serious the interference of the sensor measured value in the corresponding true abnormal area, and the more it should be corrected with higher intensity.

[0229] As an exemplary embodiment, determining the soil damage index of the true abnormal area comprises:

[0230] The first step is to calculate the average of the abnormal growth indexes of all monitoring points in the real abnormal area as the first soil mass damage factor of the real abnormal area.

[0231] Here, the first soil mass damage factor is used to at least represent the comprehensive abnormal growth of all monitoring points in the real abnormal area. The greater the first soil mass damage factor, the greater the growth of the abnormal degree in the real abnormal area at the current time.

[0232] The second step is to take the range value corresponding to the abnormal state index of all monitoring points in the real abnormal area as the second soil mass damage factor of the real abnormal area.

[0233] Here, the second soil mass damage factor is used to at least represent the instability of the abnormal degree in the real abnormal area. The greater the second soil mass damage factor, the more obvious the abnormal fluctuation in the real abnormal area at the current time.

[0234] Of course, the second soil mass damage factor of the real abnormal area can also be determined by calculating the variance or standard deviation.

[0235] The third step is to take the ratio of the area of the real abnormal area to the area of the matching abnormal area as the third soil mass damage factor of the real abnormal area.

[0236] Here, the third soil mass damage factor is used to at least represent the expansion trend of the real abnormal area. The greater the third soil mass damage factor, the greater the possibility of expansion and development of the real abnormal area at the current time.

[0237] The fourth step is to fuse the first soil mass damage factor, the second soil mass damage factor, and the third soil mass damage factor of the real abnormal area to obtain the soil mass damage index of the real abnormal area.

[0238] In one embodiment, the first soil mass damage factor, the second soil mass damage factor, and the third soil mass damage factor of the same real abnormal area are multiplied to obtain the soil mass damage index of the real abnormal area.

[0239] It should be noted that the area used to determine the third soil mass damage factor is the number of pixel points in the area, and the first soil mass damage factor, the second soil mass damage factor, and the third soil mass damage factor are all dimensionless degree values. By comprehensively analyzing the soil mass damage of the real abnormal area from different angles, the numerical accuracy of the soil mass damage index can be effectively improved.

[0240] At this point, the soil mass damage index of each real abnormal area at the current time is obtained.

[0241] S6. Based on the soil failure index, the monitoring data of each dimension at each monitoring point in each real anomaly area at the current time are corrected to obtain the corrected monitoring data set at the current time.

[0242] As an exemplary implementation, the monitoring data of each dimension at the current time corresponding to each monitoring point in each real anomaly region is corrected based on the inverse distance weight (IDW) to obtain the corrected monitoring data set at the current time, including:

[0243] For each monitoring point in each real anomaly region, the normal monitoring point closest to the target monitoring point in the real anomaly region is taken as the reference monitoring point of the target monitoring point; the IDW basic weight of the reference monitoring point is obtained based on the reference monitoring point, and the soil failure index of the real anomaly region where the target monitoring point is located is multiplied by the IDW basic weight; the normalized value of the product of the soil failure index and the IDW basic weight is used as the weight, and the IDW dynamically weights and corrects the monitoring data of each dimension corresponding to the target monitoring point at the current time.

[0244] Among them, the target monitoring point is any monitoring point in the real anomaly area, and the normal monitoring point is other monitoring points that are not located in all real anomaly areas, that is, monitoring points that do not exist at the time of real anomaly.

[0245] After obtaining the corrected monitoring data set for the current moment, the corrected monitoring data set for the current moment is output to provide real-time early warning through the output device. Specifically, the monitoring data of each sensor in different real anomaly areas is corrected through the Internet of Things platform, and the corrected monitoring data set is output. The monitoring data set includes the corrected monitoring data for each dimension at the current moment.

[0246] This invention provides a method for correcting engineering monitoring data based on an Internet of Things (IoT) data platform. This method adjusts the setting of monitoring points in different areas according to the structural changes in different areas during the actual construction of a building foundation pit, which facilitates the capture of local minor deformations. It also determines the cause of the anomaly based on the abnormal fluctuations of the data measured by each monitoring point after adjustment, thereby identifying the real sudden anomaly of the soil. Based on the deterioration trend of the real sudden anomaly area, it determines the degree of data correction required for the monitoring points therein, effectively improving the numerical accuracy of the corrected monitoring data set at the current moment.

[0247] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for correcting engineering monitoring data based on an Internet of Things (IoT) data platform, characterized in that, Includes the following steps: When inspecting the current construction site of the foundation pit, the point clouds of the foundation pit in the current period and the historical period are obtained; the historical period is the previous period of the current period, and the point clouds of the foundation pit in the current period and the historical period overlap. Based on the analysis of the physical structure complexity of local areas in the current period and historical period building foundation pit point cloud, the number of monitoring points distributed in each local area in the current period building foundation pit point cloud is determined. In each local area, the monitoring data sequence of several dimensions corresponding to the number of monitoring points distributed in the monitoring points is obtained, and the abnormal indicators of each dimension at each time are determined based on the monitoring data sequence of each dimension corresponding to each monitoring point. Based on the abnormal indicators and monitoring data of each dimension at each time, the abnormal state indicators of each monitoring point at each undetermined abnormal time are determined, and then several real abnormal times are selected based on the abnormal state indicators. Obtain each real anomaly region at each real anomaly time. Based on the anomaly status index and area of ​​each monitoring point in each real anomaly region at each real anomaly time, determine the soil failure index of each real anomaly region at the current time. Based on the soil failure index, the monitoring data of each dimension at the current time corresponding to each monitoring point in each of the real abnormal areas are corrected to obtain the corrected monitoring data set at the current time. The determination of the outlier indicators for each dimension at each time step includes: For each monitoring point, based on the monitoring data sequence of each dimension, the first-order difference sequence of each dimension is determined; Based on the first-order difference sequence of each dimension, the monitoring fluctuation characterization value of each dimension at each time moment is determined; the monitoring fluctuation characterization value is used at least to characterize the monitoring fluctuation degree at any time moment relative to historical time moments; Based on the monitoring data sequence for each dimension, determine the monitoring difference characterization value for each dimension at each time point; the monitoring difference characterization value is used at least to characterize the degree of difference in monitoring data at any given time point relative to historical time points; Obtain the time-series curve of the monitoring data sequence for each dimension, and determine the monitoring disorder characterization value of each dimension at each moment based on the fluctuation period corresponding to the adjacent extreme points on the time-series curve of each dimension. The monitoring fluctuation characterization value, the monitoring difference characterization value, and the monitoring disorder characterization value of the same dimension at the same time are fused to obtain the abnormal indicators of each dimension at each time.

2. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 1, characterized in that, Determining the distribution number of monitoring points in each local area of ​​the building foundation pit point cloud for the current period includes: The point cloud of the building foundation pit in the current period and the historical period is divided into regions to obtain the local regions corresponding to the current period and the historical period. Determine the matching region of each local region in the current cycle in the historical cycle, and then determine the vector magnitude of each point in each local region of the current cycle pointing to its matching point in the corresponding matching region; Based on the area of ​​each local region and its matching region in the current cycle, the normal vector of each point, and the vector magnitude of each point in each local region, the monitoring point distribution index of each local region in the current cycle is determined. The initial number of monitoring points in each local area of ​​the current period is adjusted using the monitoring point distribution index to obtain the number of monitoring points in each local area of ​​the building foundation pit point cloud in the current period.

3. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 2, characterized in that, The determination of the monitoring point distribution index for each local area in the current period includes: For each local region in the current cycle, the distribution factor of the first monitoring point in the local region is determined based on the area difference between the local region and its matching region. Based on the vector magnitude of each point in the local area, determine the distribution factor of the second monitoring point in the local area; The distribution factor of the third monitoring point in the local area is determined based on the similarity of the normal vectors between each point in the local area and its matching point. The distribution factors of the first, second, and third monitoring points in a local area are fused to obtain the monitoring point distribution index of the local area.

4. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 2, characterized in that, After obtaining the number of monitoring points distributed in each local area of ​​the building foundation pit point cloud for the current period, it also includes: For each local region in the current period, if the number of monitoring points in the local region is less than the number of monitoring points in the matching region, the monitoring point distribution of the overlapping region corresponding to the local region and the matching region remains unchanged, and the monitoring points of the remaining region corresponding to the local region other than the overlapping region are screened out. If the number of monitoring points in a local area is greater than the number of monitoring points in a matching area, then the distribution of monitoring points in the overlapping areas of the local area and the matching area remains unchanged, and monitoring points are added to the remaining areas of the local area other than the overlapping areas.

5. The method for correcting engineering monitoring data based on an Internet of Things (IoT) data platform according to claim 1, characterized in that, The determination of the monitoring disorder representation value for each dimension at each time step includes: For each dimension, the time interval between adjacent extreme points on the time series curve is obtained and denoted as the fluctuation period; Calculate the average of all fluctuation periods as a reference fluctuation period; Calculate the difference between the fluctuation period at each moment and the reference fluctuation period, and denot it as the difference period; The monitoring disorder characterization value for each moment is determined based on the proportion of the difference period in the reference fluctuation period.

6. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 1, characterized in that, The determination of the abnormal state indicators of each monitoring point at each pending abnormal time includes: For each monitoring point, an abnormal threshold is set. The time when the abnormal index is greater than the abnormal threshold is regarded as the time when the abnormal index is to be considered as an abnormal time, and the time other than the time when the abnormal index is to be considered as a normal time, so as to obtain each time period when the abnormal index is to be considered as an abnormal time and each time period when the abnormal index is to be considered as an abnormal time. Based on the monitoring data subsequences of each dimension in each undetermined abnormal period and each normal period, determine the comparative monitoring data subsequences of each undetermined abnormal period in each comparative dimension and the comparative monitoring data subsequences of each normal period in each comparative dimension of the target dimension; the target dimension is any dimension, and the comparative dimensions are the remaining dimensions other than the target dimension; Based on the similarity between the monitoring data subsequence of the target dimension in each undetermined anomaly period and the corresponding comparative monitoring data subsequence of the undetermined anomaly period, the first monitoring similarity index of the target dimension in each undetermined anomaly period is determined; similarly, the second monitoring similarity index of the target dimension in the normal period is determined. Based on the first and second monitoring similarity indicators for each dimension, the sequence of abnormal indicators in all pending abnormal periods, and the maximum number of abnormal dimensions at each pending abnormal time, the abnormal status indicators of the monitoring point at each pending abnormal time are determined.

7. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 6, characterized in that, The determination of the abnormal state indicators of the monitoring point at each pending abnormal time includes: Based on the differences between the first monitoring similarity indicators and the second monitoring similarity indicators in the same dimension, the first abnormal state factor of the monitoring point at each pending abnormal time is determined. Obtain the first-order difference sequence of the abnormal indicator sequence for each dimension in all undetermined abnormal periods. If the first-order difference value is not greater than zero for at least a preset number of consecutive times in the first-order difference sequence, then the corresponding time is recorded as the time when the system recovers to stability. Based on the first duration from the first pending anomaly time to the recovery stabilization time for each dimension, and the second duration from the first pending anomaly time to the current pending anomaly time, determine the second anomaly state factor of the monitoring point at each pending anomaly time; Based on the comparison between the abnormal indicators of each dimension at each pending abnormal time and the abnormal indicators of the corresponding dimension at the first pending abnormal time, the third abnormal state factor of the monitoring point at each pending abnormal time is determined. For each pending anomaly time, obtain the number of dimensions corresponding to the pending anomaly time and each pending anomaly time before the pending anomaly time, and take the maximum number of dimensions as the maximum number of anomaly dimensions at the pending anomaly time. The first, second, and third abnormal state factors and the maximum number of abnormal dimensions at the same pending abnormal time are fused to obtain the abnormal state index of the monitoring point at each pending abnormal time.

8. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 1, characterized in that, The determination of the soil failure index for each real anomaly region at the current moment includes: If the current time is a real abnormal time, then for each real abnormal region, obtain the abnormal state index sequence of the same monitoring point in the real abnormal region up to the current time, and then determine the first difference sequence of the abnormal state index sequence. Based on the first-order difference sequence of each monitoring point in the real anomaly region at the current moment, determine the abnormal growth index of each monitoring point in the real anomaly region; The real anomaly region at the current moment is matched with each real anomaly region at the first real anomaly moment to obtain the matched anomaly region of the real anomaly region at the current moment; Based on the abnormal growth index of each monitoring point in the real abnormal area, the range of the abnormal state indicators of all monitoring points, and the area of ​​the real abnormal area and the matching abnormal area, the soil failure index of the real abnormal area is determined.

9. The method for correcting engineering monitoring data based on an Internet of Things data platform according to claim 8, characterized in that, The determination of the soil failure index in the actual abnormal area includes: The average value of the abnormal growth index of all monitoring points in the real abnormal area is calculated as the first soil failure factor in the real abnormal area. The range of abnormal state indicators of all monitoring points in the real abnormal area is used as the second soil failure factor of the real abnormal area. The ratio of the area of ​​the real anomaly region to the area of ​​the matched anomaly region is used as the third soil failure factor of the real anomaly region. The first, second, and third soil failure factors of the real anomaly area are fused to obtain the soil failure index of the real anomaly area.

Images