Deep and large foundation pit deformation monitoring method and system based on multi-modal data fusion

Abstract

The invention discloses a deep and large foundation pit deformation monitoring method based on multi-modal data fusion, and the method comprises the steps: monitoring the natural environment of a target foundation pit, obtaining monitoring data, and obtaining an initial data monitoring set of the target foundation pit; according to the initial data monitoring set of the target foundation pit, constructing a foundation pit state description model for describing the target foundation pit to obtain a collapse risk value of a corresponding foundation pit collapse test; performing data reprocessing according to the collapse risk value, and constructing a collapse risk influence parameter relationship network for describing the foundation pit collapse test; performing target foundation pit collapse risk analysis and assessment through the collapse risk description graph to obtain a corresponding collapse risk assessment description; and a collapse risk warning model is set according to the collapse risk assessment description, and deformation monitoring and early warning are conducted on the target foundation pit. According to the scheme, through multi-parameter dynamic trend analysis, a collaborative change rule among different parameters is captured, a multi-scene test is carried out by constructing a three-dimensional model, and key risk factor combinations in different construction stages can be identified.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of image processing, and particularly relates to a deep and large foundation pit deformation monitoring method and system based on multi-modal data fusion. BACKGROUND

[0002] With the acceleration of urbanization and the development and utilization of underground space, large deep foundation pit projects are increasing. These projects often have characteristics such as large excavation depth, complex surrounding environment, and variable geological conditions, which pose unprecedented challenges to monitoring technology.

[0003] At present, some foundation pit monitoring systems install strain gauges and other equipment on the surface of the supporting structure to measure the stress state of the supporting structure and determine whether it exceeds the warning value. However, the data obtained by the related monitoring method is limited and lacks multi-dimensional description of the foundation pit environment. The single type of data may lead to incomplete judgment of potential risks.

[0004] Therefore, it is urgent to propose a deep and large foundation pit deformation monitoring method and system based on multi-modal data fusion to solve the above problems. SUMMARY

[0005] To solve the above problems in the prior art, the purpose of the present application is to overcome the existing deficiencies and provide a deep and large foundation pit deformation monitoring method based on multi-modal data fusion, characterized by comprising: Monitoring the natural environment of the target foundation pit to obtain monitoring data, and performing data processing on the obtained multi-modal data to obtain an initial data monitoring set of the target foundation pit; Constructing a foundation pit state description model for describing the target foundation pit according to the initial data monitoring set of the target foundation pit, performing a foundation pit collapse test based on the constructed foundation pit state description model, obtaining test collapse detection values formed after the foundation pit collapse test, and obtaining a collapse risk value corresponding to the foundation pit collapse test through the test collapse detection values; According to the collapse risk value, data is reprocessed to construct a collapse risk influence parameter relationship network for describing the foundation pit collapse test; the target foundation pit is calibrated and rated for collapse risk based on the initial data monitoring set, and a corresponding collapse risk description graph is drawn; Performing collapse risk analysis and evaluation of the target foundation pit through the collapse risk description graph to obtain a corresponding collapse risk evaluation description; setting a collapse risk warning model according to the collapse risk evaluation description, setting a corresponding collapse risk warning line, and performing deformation monitoring and early warning of the target foundation pit based on the risk warning line.

[0006] As a further optimization of the above scheme, the generation step of the collapse risk influence parameter relationship network comprises the following: Based on the historical collapse record of the target foundation pit and the initial data monitoring set, the target to be analyzed is obtained, and the parameters to be analyzed of the corresponding type of target foundation pit are obtained; According to the collapse risk value of the foundation pit collapse test, the correlation analysis of the parameters to be analyzed of the target foundation pit and the historical collapse record of the corresponding target foundation pit is carried out, and the correlation degree of the parameters to be analyzed of the target foundation pit and the historical collapse record of the corresponding target foundation pit is obtained, and then the relationship between the parameters to be analyzed of the target foundation pit and the historical collapse record of the corresponding target foundation pit is calculated. According to the calculated relationship, the key influence analysis of the parameters to be analyzed of the target foundation pit is carried out, and the influence degree of the parameters to be analyzed of the target foundation pit on the historical collapse record of the target foundation pit is obtained; according to the calculated influence degree, the parameters to be analyzed of the target foundation pit are divided into high-risk influence factors and ordinary influence factors; the relationship architecture of high-risk influence factors and ordinary influence factors is established; According to the divided high-risk influence factors and ordinary influence factors, 2 factor combinations are randomly selected as an analysis group, and a plurality of analysis groups are randomly generated; The mutual relationship promoting analysis is carried out for the generated plurality of analysis groups, the corresponding mutual relationship promoting description is obtained, and the combination frequency and risk frequency of the plurality of analysis groups are recorded; According to the historical collapse record of the target foundation pit, the time sequence relationship of the high-risk influence factors and the ordinary influence factors is calculated. According to the influence degree, the relationship architecture, the mutual relationship promoting description and the time sequence relationship, the collapse risk influence parameter relationship network for describing the target foundation pit is generated.

[0007] As a further optimization of the above scheme, the initial data monitoring set includes data acquisition time and foundation pit acquisition engineering parameters.

[0008] As a further optimization of the above scheme, the calculation method of the collapse risk value includes the following steps: A 3D model is constructed for the initial data monitoring set, and a foundation pit state description model for describing the target foundation pit is constructed; The historical collapse record of the target foundation pit is obtained; According to the constructed foundation pit state description model, a foundation pit collapse test is carried out, and test collapse detection values of a plurality of collapse conditions are obtained; According to any test collapse detection value of the collapse condition, all detection parameters of the test collapse detection value of this round are counted, and a trend graph of the test collapse detection value is drawn with the foundation pit collapse test time as the x-axis and any detection parameter as the y-axis. The image features of the drawn trend graph are obtained, and the image features are used as the collapse risk value of this round test; Repeat the above steps until the collapse risk value calculation and statistics of the plurality of collapse conditions are completed.

[0009] As a further optimization of the above scheme, the drawing step of the collapse risk description diagram comprises the following: According to the constructed collapse risk influence parameter relationship network and the foundation pit state description model, a plurality of calibration regions are obtained for the target foundation pit; According to the obtained plurality of calibration regions, stress detection is performed to obtain stress detection parameters of the plurality of calibration regions; high-risk influence factors and ordinary influence factors and stress detection parameters are constructed together to form a calibration region stability model, and the output end of the calibration region stability model is the collapse probability of the plurality of calibration regions entered; According to the output plurality of collapse probabilities, collapse risk part calibration and collapse risk part rating are performed to draw the corresponding collapse risk description diagram.

[0010] As a further optimization of the above scheme, the obtaining of the collapse risk assessment description specifically comprises the following: According to the obtained initial data monitoring set, data collection time statistics and target foundation pit trend change records of the initial data monitoring set are performed to obtain a first system for recording the initial data monitoring set about the collection time and trend change records; According to the collapse risk description diagram, a foundation pit space correlation model and a foundation pit space relationship network of the target foundation pit are established, and a target foundation pit collapse test is carried out according to the constructed foundation pit space relationship network, and a collapse point is recorded to obtain a second system; The first system and the second system are integrated to obtain a development evaluation of the target foundation pit; According to the development evaluation, target foundation pit collapse deformation monitoring is performed on the foundation pit to be monitored, and a confidence level is obtained according to the collapse deformation monitoring result, and an interval value where the confidence level is located is calculated; According to the obtained interval value where the confidence level is located and the corresponding collapse deformation monitoring result, data integration is performed to obtain the corresponding collapse risk assessment description.

[0011] As a further optimization of the above scheme, the obtaining step of the collapse risk alarm line specifically comprises the following: According to the collected collapse risk assessment description, the high-risk influence factors and the ordinary influence factors are optimized to obtain a mutual relationship promotion description of the optimized high-risk influence factors and the ordinary influence factors, and a collapse risk alarm model is established according to the mutual relationship promotion description; Based on the collapse risk alarm model, initial data distribution analysis and initial collapse risk alarm line analysis of the target foundation pit are performed; According to the historical collapse record of the target foundation pit, the result of the initial collapse risk warning line analysis is evaluated, and whether to optimize the initial collapse risk warning line analysis is determined based on the evaluation result of the initial collapse risk warning line analysis: If optimization is not needed, the initial collapse risk warning line analysis is directly output as the final collapse risk warning line; If optimization is needed, the initial collapse risk warning line analysis is optimized based on an optimization method.

[0012] As a further optimization of the above scheme, the step of deforming the target foundation pit based on the risk warning line and warning includes the following: According to the collected initial data monitoring set, data feature analysis is performed to obtain the data features actually corresponding to the high-risk influence factors and ordinary influence factors; The data features actually corresponding to the high-risk influence factors and ordinary influence factors are matched with the collapse risk warning line, and instructions for deforming the target foundation pit and warning are output according to the matching result.

[0013] A computer program product comprising instructions, characterized in that when the computer program product is run on a deep foundation pit monitoring system, the deep foundation pit deformation monitoring system based on multi-modal data fusion executes the method of any of the claims.

[0014] A computer-readable storage medium comprising instructions, characterized in that when the instructions are run on a deep foundation pit monitoring system, the deep foundation pit deformation monitoring system based on multi-modal data fusion executes the method of any of the claims.

[0015] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. The present scheme can capture the cooperative change law between different parameters through multi-parameter dynamic trend analysis, and can identify key risk factor combinations in different construction stages by constructing a three-dimensional model and carrying out multi-scenario testing.

[0016] 2. The present scheme realizes risk decomposition in three-dimensional space through modular calibration, improves the spatio-temporal resolution of risk prediction by combining time series data analysis models, and realizes fine spatial classification of foundation pit collapse risk through the above technical solutions, so that the risk degree of different regions is intuitively presented through color coding. The introduction of dynamic stress parameters enhances the real-time performance of risk assessment, and can capture abnormal changes in local areas in time. BRIEF DESCRIPTION OF DRAWINGS

[0017] Other features, objects and advantages of the present application will become more apparent through reading the detailed description of the non-limiting embodiments made by referring to the following drawings: Figure 1 A flowchart is provided for the present application. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are merely exemplary for the purpose of explanation and are not intended to limit the application.

[0019] As shown in the drawings, Figure 1 The embodiment of the present application discloses a deep foundation pit deformation monitoring method based on multi-modal data fusion, which specifically comprises the following steps: The natural environment of the target foundation pit is monitored, and monitoring data is obtained. The obtained multi-modal data is processed to obtain an initial data monitoring set of the target foundation pit. It should be particularly noted that the natural environment monitoring mentioned in the present application includes parameters such as the annual climate of the target foundation pit and the regional soil quality. The data processing mentioned in the present application specifically represents data feature processing and extraction based on the obtained multi-modal data, and data fusion is performed according to the processed data, so as to obtain an effective data set that can directly describe the target foundation pit. A foundation pit state description model for describing the target foundation pit is constructed according to the initial data monitoring set of the target foundation pit. A foundation pit collapse test is performed based on the constructed foundation pit state description model, and a test collapse detection value formed after the foundation pit collapse test is obtained. A collapse risk value corresponding to the foundation pit collapse test is obtained through the test collapse detection value. More specifically, the test collapse detection value represents a simulation collapse test on the target foundation pit, and further obtained collapse process monitoring parameters. According to the collapse risk value, data reprocessing is performed to construct a collapse risk influence parameter relationship network for describing the foundation pit collapse test. The target foundation pit is calibrated and rated according to the initial data monitoring set, and a corresponding collapse risk description graph is drawn. More specifically, the specific collapse risk Through the collapse risk description graph, the collapse risk of the target foundation pit is analyzed and evaluated, and a corresponding collapse risk evaluation description is obtained. According to the collapse risk evaluation description, a collapse risk warning model is set, a collapse risk warning line is set, and deformation monitoring and early warning of the target foundation pit are performed based on the risk warning line.

[0020] More specifically, the acquisition of multi-modal data includes automatic real-time collection: embedding or installing various devices such as strain gauges, inclinometers, static water levels, etc. at key positions of the foundation pit to realize the acquisition of the mechanical properties and structural response data of the foundation pit; for monitoring items such as ground settlement and support pile top displacement, survey personnel use precision levels, total stations and other equipment to conduct field observation and record data; high-definition video cameras are used for continuous video monitoring, and ground interferometric radar is used to scan along the perimeter of the foundation pit to obtain data, and meteorological data of the target foundation pit are continuously obtained; through the information collection of the above technical means, the mechanical properties and structural response data of the foundation pit, field observation data of the foundation pit, radar scanning data and meteorological data are fused to form multi-modal data. Due to the time or geographical location difference in the multi-modal data acquisition process of the present application, there may be obviously abnormal collected data; for abnormal data, a unified time standard, geographical position effective range value constraint is used, and professional technical personnel are used for judgment and screening, so that in the subsequent calculation process, a unified time record value, plane coordinate, spatial coordinate and other geographical position description are used.

[0021] More specifically, the present application further analyzes and integrates the screened multi-modal data to form available initial parameter description values that can be used to describe the current state or future deformation of the target foundation pit, which can be divided into collection time of foundation pit data, geographical collection difference of foundation pit data and construction parameters of different foundation pits.

[0022] Specifically, the generation steps of the collapse risk influence parameter relationship network include the following: Based on the historical collapse records of the target foundation pit and the initial data monitoring set, the target to be analyzed is obtained, and the parameters to be analyzed of the corresponding type of target foundation pit are obtained; it should be particularly noted that the historical collapse record of the present application specifically represents: the accident process parameter record that can be used to describe or record the deformation and collapse of the foundation pit, including but not limited to the current foundation pit to be monitored and processed, and further, the foundation pit with the same soil layer, the same purpose and the same scale can be regarded as the preferred historical collapse record of any selected target foundation pit; based on the actual construction requirements, the "same" mentioned in the present application can be regarded as approximate, and the specific construction personnel statistics is used as the criterion; According to the collapse risk value of the foundation pit collapse test, the correlation between the parameters to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit is analyzed, the correlation degree between the parameters to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit is obtained, and the relationship between the parameters to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit is calculated; According to the calculated relationship, the key influence analysis of the to-be-analyzed parameters of the target foundation pit is performed, and the influence degree of the to-be-analyzed parameters of the target foundation pit on the historical collapse record of the target foundation pit is obtained; according to the calculated influence degree, the to-be-analyzed parameters of the target foundation pit are divided into high-risk influence factors and ordinary influence factors; and a relationship framework of the high-risk influence factors and the ordinary influence factors is established; According to the divided high-risk influence factors and ordinary influence factors, 2 factor combinations are randomly selected as an analysis group, and a plurality of analysis groups are randomly generated; The mutual relationship promotion analysis is performed on the generated plurality of analysis groups, and the corresponding mutual relationship promotion description is obtained, and the combination frequency and the risk frequency of the plurality of analysis groups are recorded; more specifically, the combination frequency is characterized in the present application as the probability of the occurrence of the analysis group of the corresponding high-risk influence factor or ordinary influence factor in the collapse test according to the record data in the historical collapse record; the risk frequency is characterized as the condition change of the occurrence of the analysis group of the corresponding high-risk influence factor or ordinary influence factor, and specifically, the building foundation pit is more prone to collapse under the combined action of high humidity and building applied force on the edge of the foundation pit, the combination frequency is represented as the frequency of the combined occurrence of high humidity and building applied force on the edge of the foundation pit, and the risk frequency is the condition of the increase or decrease of the collapse; The time sequence relationship of the high-risk influence factors and the ordinary influence factors is calculated according to the historical collapse record of the target foundation pit; According to the influence degree, the relationship framework, the mutual relationship promotion description and the time sequence relationship, a collapse risk influence parameter relationship network for describing the target foundation pit is generated.

[0023] Further, the support level of the present application refers to the probability of the co-occurrence of a specific parameter and a collapse event, which can be realized by using the support level calculation formula in association rule mining, and is used for screening parameters strongly related to collapse; the confidence level refers to the risk frequency of the occurrence of a collapse event under the premise of the occurrence of a parameter, which can be calculated by using the Bayes theorem, and is used for verifying the causal relationship between the parameter and the collapse; the influence degree refers to the influence weight of the parameter on the collapse result, and is used for distinguishing between key parameters and non-key parameters; the time sequence relationship refers to the correlation mode of the parameter change with time, which can be realized by using time series cross-correlation analysis or dynamic time warping algorithm, and is used for capturing the lag effect of the parameter change; the association degree of the to-be-analyzed parameters of the target foundation pit and the corresponding historical collapse record of the target foundation pit can be represented by using the confidence level and the support level; Specifically, first, geological parameters, support structure stress values, groundwater levels, and other monitoring data are extracted from historical collapse records as a set of parameters to be analyzed. By calculating the support level of each parameter in the collapse event, for example, when the support structure displacement exceeds the threshold value, the support level of the collapse is 0.85, high-risk parameters such as insufficient support and high soil moisture content are screened out. The random forest model is used to evaluate the influence degree of each parameter, and the influence degree higher than the set threshold (for example, 0.7) is divided into high-risk factors. The relationship framework is established between the high-risk factors and the ordinary factors, for example, the geological conditions are taken as the first-level factors, and the construction parameters are taken as the second-level factors. Two factors are randomly selected to form an analysis group, for example, the combination of soil moisture content and support structure displacement, the combination frequency is 0.68, and the risk frequency is 0.79. The synergistic effect significantly increases the collapse risk. Through time series analysis, it is found that the groundwater level change is 3 days ahead of the support displacement change, and the time sequence relationship is established. Finally, all the analysis results are integrated to construct a multi-dimensional relationship network including parameter correlation strength, synergistic effect and time sequence characteristics.

[0024] Through the above technical scheme, the present scheme realizes dynamic correlation modeling of multi-source parameters, can accurately identify high-risk influence factors and their interaction mechanism, and the parameter relationship network established can quantify the synergistic effect of multiple factors. Through objective influence degree division, human error is avoided, such as accurately identifying the influence degree of neglected geological structure parameters reaching 0.82, and reducing the false negative rate by 35% after including the high-risk factors. The finally formed multi-dimensional relationship network provides comprehensive data support for risk assessment, and the collapse monitoring accuracy reaches more than 93%.

[0025] Specifically, the initial data monitoring set includes data collection time and foundation pit collection engineering parameters.

[0026] Specifically, the calculation method of the collapse risk value includes the following steps: A 3D model is constructed for the initial data monitoring set to construct a foundation pit state description model for describing the target foundation pit; Obtain the historical collapse record of the target foundation pit; According to the constructed foundation pit state description model, a foundation pit collapse test is performed to obtain test collapse detection values of multiple collapse conditions; According to any test collapse detection value of the collapse condition, all detection parameters of the test collapse detection value of this round are counted, taking the foundation pit collapse test time as the x-axis and any detection parameter as the y-axis, a trend graph of the test collapse detection value is drawn; the image features of the drawn trend graph are obtained, and the image features are taken as the collapse risk value of this round test; Repeat the above steps until the collapse risk value calculation and statistics of multiple collapse conditions are completed.

[0027] Further, the foundation pit state description model refers to a three-dimensional visualization model constructed based on multi-modal monitoring data, which can be realized by BIM modeling technology or finite element analysis software, and is used to dynamically reflect the coupling relationship between the foundation pit structure state and external environmental parameters. The historical collapse record refers to the collapse event data of the target foundation pit in the region or under similar geological conditions, which can be stored in a structured database and includes time, location, and cause parameters, etc. to provide a real scene reference for the test. The trend chart refers to a dynamic change curve with time series as the horizontal axis and detection parameters as the vertical axis.

[0028] Further, a three-dimensional foundation pit model is constructed by fusing multi-modal monitoring data, and parameters such as soil pressure, displacement, and groundwater level are mapped to the spatial dimension to establish physical correlations between parameters. Based on the threshold values of key parameters of historical collapse events, different working conditions of the model are set for collapse simulation tests, such as groundwater level rising or support structure failure scenarios. In each test, the numerical sequence of each detection parameter changing with time is recorded synchronously to generate a trend chart containing characteristics such as slope, peak value, and fluctuation frequency. Key feature vectors in the trend chart are extracted by image processing algorithms and encoded into numerical indicators representing the current test collapse risk level. The simulation tests of different collapse cause combinations are repeatedly executed to finally form a set of collapse risk values covering multiple potential risk scenarios. This scheme can capture the relationship between different parameters through multi-parameter dynamic trend analysis, and can identify key risk factor combinations in different construction stages by constructing a three-dimensional model for multi-scenario testing. Specifically, the drawing steps of the collapse risk description chart include the following: According to the constructed collapse risk influence parameter relationship network and the foundation pit state description model, the target foundation pit is calibrated to obtain multiple calibration areas; Stress detection is performed on the obtained multiple calibration areas to obtain stress detection parameters of the multiple calibration areas. High-risk influence factors, ordinary influence factors, and stress detection parameters are constructed together to form a calibration area stability model. The output end of the calibration area stability model is the collapse probability of the input multiple calibration areas. According to the output multiple collapse probabilities, the collapse risk part is calibrated and rated, and the corresponding collapse risk description chart is drawn.

[0029] Further, the foundation pit state description model and the relationship network between the foundation pit state description model and the collapse risk influence parameters are input into the spatial analysis system. First, the foundation pit is modularized and divided by a geographic information system, for example, the foundation pit is divided into 3m*3m plane units. High-risk influence factors and ordinary influence factors such as soil water content data are input into the stability model at the same time, and the collapse probability of each unit is calculated by a weighted fusion algorithm. The temporal and spatial resolution of risk prediction is improved by combining the time series data analysis model. Through the above technical solutions, the fine spatial classification of the foundation pit collapse risk is realized, and the risk degree of different regions is intuitively presented through drawing, and abnormal changes in local areas can be captured in time.

[0030] Specifically, the acquisition of the collapse risk assessment description specifically includes the following: According to the initial data monitoring set, the data acquisition time is counted, and the target foundation pit trend change record of the initial data monitoring set is recorded, to obtain a first system for recording the initial data monitoring set about the acquisition time and the trend change record; According to the collapse risk description diagram, a foundation pit spatial correlation model and a foundation pit spatial relationship network of the target foundation pit are established, a foundation pit collapse test of the target foundation pit is carried out according to the established foundation pit spatial relationship network, and a collapse point is recorded, to correspondingly obtain a second system; More specifically, the foundation pit spatial correlation model specifically represents the dependency of high-risk influence factors and ordinary influence factors on the spatial structure of the target foundation pit, and can be obtained according to existing spatial analysis methods; the foundation pit spatial relationship network specifically represents the mutual spatial correlation of the plurality of calibration regions of the target foundation pit; based on the above description, the second system can realize a plurality of evaluations of the target foundation pit, such as collapse point estimation, collapse path estimation, and collapse influence range estimation; The first system and the second system are integrated and processed to obtain a development evaluation of the target foundation pit; According to the development evaluation, a target foundation pit collapse deformation monitoring of a to-be-monitored foundation pit is carried out, a confidence level is obtained according to a collapse deformation monitoring result, and an interval value where the confidence level is located is calculated; According to the interval value where the confidence level is located and the corresponding collapse deformation monitoring result, data integration is performed to obtain a corresponding collapse risk assessment description.

[0031] Further, time series data sets are formed by continuously collecting foundation pit displacement and support shaft force parameters, a first system containing hourly change rate and daily fluctuation amplitude characteristics is established, and the evolution trajectory of the foundation pit state is tracked in real time. Based on the collapse risk description diagram, the spatial coordinates and risk levels of each monitoring point are extracted, a spatial network model containing adjacent point correlation weights and risk diffusion coefficients is constructed, the trend anomaly in the time dimension and the conduction effect in the space dimension are coupled and analyzed, and a high-risk area with spatiotemporal correlation is identified as a development evaluation result.

[0032] Specifically, the obtaining step of the collapse risk warning line specifically comprises the following: According to the collected collapse risk assessment description, the high-risk influence factors and the general influence factors are optimized to obtain the mutual relationship promotion description of the optimized high-risk influence factors and the general influence factors, and a collapse risk warning model is established according to the mutual relationship promotion description; Based on the collapse risk warning model, initial data distribution analysis and initial collapse risk warning line analysis of the target foundation pit are performed; According to the historical collapse record of the target foundation pit, the result of the initial collapse risk warning line analysis is evaluated, and based on the evaluation result of the initial collapse risk warning line analysis, it is determined whether to optimize the initial collapse risk warning line analysis: If optimization is not needed, the initial collapse risk warning line analysis is directly output as the final collapse risk warning line; If optimization is needed, the initial collapse risk warning line analysis is optimized based on an optimization method.

[0033] Further, the correlation between the influence factors is dynamically corrected through the collapse risk assessment description, so that the warning model can adapt to the real-time changes of the foundation pit state. In the initial warning line generation stage, a statistical benchmark is established in combination with the data distribution characteristics to ensure that the warning threshold matches the current monitoring data characteristics. The initial warning line is verified through the historical collapse record. When the actual triggering condition of the collapse event in the historical data deviates from the warning line, the optimization algorithm is started to adjust the warning threshold parameter.

[0034] Through the above technical solution, the accuracy and timeliness of risk early warning in foundation pit deformation monitoring are improved. Through the dynamic optimization mechanism, the warning threshold can reflect the changes of the foundation pit state in real time, reducing the risk of false positives. Combined with the evaluation and verification process of the historical collapse record, the scientificity and reliability of the warning line are ensured, and the warning system formed can timely capture the collapse risk evolution trend, providing precise early warning support for engineering safety.

[0035] Specifically, the steps of performing deformation monitoring and early warning on the target foundation pit based on the risk warning line specifically comprise the following: According to the collected initial data monitoring set, data feature analysis is performed to obtain the data features actually corresponding to the high-risk influence factors and the general influence factors; The data features actually corresponding to the high-risk influence factors and the general influence factors are matched with the collapse risk warning line, and instructions for performing deformation monitoring and early warning on the target foundation pit are output according to the matching result.

[0036] Further, the multi-source information of the initial monitoring data is converted into a standardized feature vector through data feature analysis, the influence of different dimensional data on the analysis result is eliminated, the feature vector is input into a pre-established alarm line model, and the deviation degree of the current feature from the safety threshold is compared in real time through a sliding window mechanism; when the feature vector breaks through the alarm line, a pre-warning instruction including a risk position, a risk level and a disposal suggestion is automatically generated by the system and directly pushed to the monitoring terminal; the comparability between different monitoring parameters is improved by converting the multi-modal data into a unified feature space; and the adaptability of the system to the change of the construction environment is enhanced by establishing a dynamic alarm line model.

[0037] It should be noted that the above-mentioned embodiment of the present application is a deep and large foundation pit deformation monitoring system based on multi-modal data fusion, and the implementation technical means thereof is the same as that of a deep and large foundation pit deformation monitoring method based on multi-modal data fusion, and thus is not described in detail here.

[0038] It should be noted that in this paper, the term "including" "containing" or any other variant thereof is intended to cover non-exclusive containing, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitation, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the method and device in the embodiments of the present application is not limited to performing the functions in the order shown or discussed, but can also include performing the functions in a substantially simultaneous manner or in a reverse order according to the functions involved, for example, the described method can be performed in an order different from that described, and various steps can also be added, omitted or combined. In addition, the features described with reference to certain examples can be combined in other examples.

[0039] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment method can be realized by means of software and a necessary general hardware platform, of course, it can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disc, optical disc), including a plurality of instructions, for making a terminal (which can be a mobile phone, computer, server, or network device, etc.) execute the method of each embodiment of the present application.

[0040] The embodiments of the present application are described above with reference to the drawings, but the present application is not limited to the specific embodiments described above, and the specific embodiments described above are merely illustrative, not restrictive, and a person of ordinary skill in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims, all of which belong to the protection of the present application.

[0041] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an illustrative embodiment", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0042] Although the embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for deep foundation pit deformation monitoring based on multi-modal data fusion, characterized in that, The method comprises the following steps: monitoring the natural environment of a target foundation pit, obtaining monitoring data, processing the obtained multi-modal data to obtain an initial data monitoring set of the target foundation pit; constructing a foundation pit state description model for describing the target foundation pit according to the initial data monitoring set of the target foundation pit, performing a foundation pit collapse test based on the constructed foundation pit state description model, obtaining test collapse detection values formed after the foundation pit collapse test, and obtaining a collapse risk value corresponding to the foundation pit collapse test through the test collapse detection values; performing data reprocessing according to the collapse risk value, constructing a collapse risk influence parameter relationship network for describing the foundation pit collapse test, calibrating and rating a collapse risk part of the target foundation pit based on the initial data monitoring set, and drawing a corresponding collapse risk description graph; performing collapse risk analysis and evaluation of the target foundation pit through the collapse risk description graph, and obtaining a corresponding collapse risk evaluation description; setting a collapse risk warning model according to the collapse risk evaluation description, setting a corresponding collapse risk warning line, and performing deformation monitoring and early warning of the target foundation pit based on the risk warning line.

2. The deep foundation pit deformation monitoring method based on multi-modal data fusion according to claim 1, characterized in that, The generation step of the collapse risk influence parameter relationship network comprises the following steps: obtaining a target to be analyzed based on historical collapse records of the target foundation pit and the initial data monitoring set, and obtaining a target foundation pit of a corresponding type and a parameter to be analyzed; performing correlation analysis on the parameter to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit according to the collapse risk value of the foundation pit collapse test, obtaining the correlation degree of the parameter to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit, and then calculating the relationship between the parameter to be analyzed of the target foundation pit and the historical collapse records of the corresponding target foundation pit; performing key influence analysis on the parameter to be analyzed of the target foundation pit according to the calculated relationship, obtaining the influence degree of the parameter to be analyzed of the target foundation pit on the historical collapse records of the target foundation pit, dividing the parameter to be analyzed of the target foundation pit into a high-risk influence factor and a common influence factor according to the calculated influence degree, and establishing a relationship framework of the high-risk influence factor and the common influence factor; arbitrarily selecting two factor combinations as an analysis group according to the divided high-risk influence factor and common influence factor, and randomly generating multiple analysis groups; performing mutual relationship promotion analysis on the generated multiple analysis groups, obtaining a corresponding mutual relationship promotion description, and recording the combination frequency and risk frequency of the multiple analysis groups; calculating the time sequence relationship of the high-risk influence factor and the common influence factor according to the historical collapse records of the target foundation pit; generating a collapse risk influence parameter relationship network for describing the target foundation pit according to the influence degree, the relationship framework, the mutual relationship promotion description, and the time sequence relationship.

3. The deep foundation pit deformation monitoring method based on multi-modal data fusion according to claim 2, characterized in that, The initial data monitoring set comprises data acquisition time and foundation pit acquisition engineering parameters.

4. The deep foundation pit deformation monitoring method based on multi-modal data fusion according to claim 3, characterized in that, The calculation method of the collapse risk value comprises the following steps: constructing a 3D model for the initial data monitoring set, and constructing a foundation pit state description model for describing the target foundation pit; obtaining historical collapse records of the target foundation pit; performing a foundation pit collapse test according to the constructed foundation pit state description model, and obtaining test collapse detection values of multiple collapse conditions; According to the test collapse detection value of any one of the collapse conditions, all detection parameters of the test collapse detection value of this round of test are counted, a trend graph of the test collapse detection value is drawn with the foundation pit collapse test time as the x-axis and any one of the detection parameters as the y-axis, and an image feature of the drawn trend graph is obtained, which is taken as the collapse risk value of this round of test; The above steps are repeated until the collapse risk value calculation and statistics of multiple collapse conditions are completed.

5. The deep foundation deformation monitoring method based on multi-modal data fusion according to claim 4, characterized in that, The drawing step of the collapse risk description graph includes the following: According to the constructed collapse risk influence parameter relationship network and the foundation pit state description model, module calibration is performed for the target foundation pit to obtain multiple calibration regions; Stress detection is performed on the obtained multiple calibration regions to obtain stress detection parameters of the multiple calibration regions; high-risk influence factors, ordinary influence factors, and stress detection parameters are combined to construct a calibration region stability model, and an output end of the calibration region stability model is a collapse probability of the multiple calibration regions; According to the output multiple collapse probabilities, collapse risk part calibration and collapse risk part rating are performed, and a corresponding collapse risk description graph is drawn.

6. The deep foundation pit deformation monitoring method based on multi-modal data fusion according to claim 5, characterized in that, The acquisition of the collapse risk assessment description specifically includes the following: According to the obtained initial data monitoring set, data acquisition time statistics are performed, and target foundation pit trend change records of the initial data monitoring set are recorded to obtain a first system for recording the initial data monitoring set of multiple states about acquisition time and trend change records; According to the collapse risk description graph, a foundation pit space correlation model and a foundation pit space relationship network of the target foundation pit are established, a target foundation pit collapse test is carried out according to the constructed foundation pit space relationship network, a collapse point is recorded, and a second system is correspondingly obtained; The constructed first system and second system are integrated to obtain a development evaluation of the target foundation pit; According to the development evaluation, target foundation pit collapse deformation monitoring is performed on the foundation pit to be monitored, a confidence level is obtained according to the collapse deformation monitoring result, and an interval value where the confidence level is located is calculated; According to the obtained interval value where the confidence level is located and the corresponding collapse deformation monitoring result, data integration is performed to obtain a corresponding collapse risk assessment description.

7. The deep foundation deformation monitoring method based on multi-modal data fusion according to claim 6, characterized in that, The acquisition step of the collapse risk alarm line specifically includes the following: According to the collected collapse risk assessment description, the high-risk influence factors and the ordinary influence factors are optimized to obtain a mutual relationship promotion description of the optimized high-risk influence factors and the ordinary influence factors, and a collapse risk alarm model is established according to the mutual relationship promotion description; Based on the collapse risk alarm model, initial data distribution analysis and initial collapse risk alarm line analysis of the target foundation pit are performed; According to the historical collapse records of the target foundation pit, the result of the initial collapse risk alarm line analysis is evaluated, and whether to optimize the initial collapse risk alarm line analysis is confirmed based on the evaluation result of the initial collapse risk alarm line analysis: If optimization is not needed, the initial collapse risk alarm line analysis is directly output as the final collapse risk alarm line; If optimization is needed, the initial collapse risk alarm line analysis is optimized based on an optimization method.

8. The deep foundation pit deformation monitoring method based on multi-modal data fusion according to claim 7, characterized in that, The step of monitoring and warning the target foundation pit based on the risk warning line specifically comprises the following steps: According to the initial data monitoring set, data feature analysis is performed to obtain data features corresponding to the high-risk influence factor and the ordinary influence factor; The data features corresponding to the high-risk influence factor and the ordinary influence factor are matched with the collapse risk warning line, and instructions for monitoring and warning the target foundation pit are output according to the matching result.

9. A computer program product comprising instructions, characterized in that, When the computer program product runs on the deep foundation pit monitoring system, the deep foundation pit deformation monitoring system based on multi-modal data fusion executes the method as claimed in any one of claims 1-8.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions run on the deep foundation pit monitoring system, the deep foundation pit deformation monitoring system based on multi-modal data fusion executes the method as claimed in any one of claims 1-8.

Images