A foundation pit deformation monitoring method and system applying image analysis

Abstract

The application discloses a kind of foundation pit deformation monitoring method and system of application image analysis, it is related to image analysis technical field, including according to deformation main body, all deformation content is divided into direct deformation content and indirect deformation content, so as to analyze the deformation correlation between different deformation main body each other, provide reliable basis for the deformation situation of subsequent engineering construction stage.Direct deformation content and indirect deformation content are combined to determine the deformation degree of foundation pit related engineering construction stage, comprehensively describe all deformation conditions about foundation pit in engineering construction stage.The deformation correlation between engineering construction stage each other is determined by the deformation correlation between all deformation main body under engineering construction stage, so as to consider the deformation influence of preceding engineering construction stage on subsequent engineering construction stage, set the threshold of deformation degree, improve the accuracy and adaptability of foundation pit deformation monitoring, ensure that foundation pit engineering subsequent management and security prevention.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image analysis, in particular to a foundation pit deformation monitoring method and system applying image analysis. BACKGROUND

[0002] With the continuous progress of artificial intelligence technology, image analysis technology has been widely applied in engineering monitoring field, especially in foundation pit deformation monitoring, which shows significant advantages. This technology acquires image data of foundation pit and its surrounding environment in real time through image acquisition devices such as cameras, and then uses advanced image processing and analysis algorithms such as edge detection, feature extraction and matching to accurately identify and quantify the deformation information in the image. Compared with traditional monitoring methods, image analysis technology has the advantages of non-contact, high precision, real-time monitoring, etc., which can comprehensively and intuitively reflect the deformation of foundation pit and provide strong guarantee for engineering safety. In recent years, with the introduction of multi-source image fusion, deep learning and other technologies, the application prospect of image analysis in foundation pit deformation monitoring is more broad.

[0003] In the prior art, for foundation pit deformation monitoring, only the current state of the foundation pit is monitored, without considering the influence of the previous engineering construction stage, which makes the precision and adaptability of the foundation pit deformation monitoring low, the reliability poor, and the current deformation of the foundation pit cannot be effectively monitored, which is not conducive to the safety of subsequent engineering.

[0004] Therefore, how to improve the precision and adaptability of foundation pit deformation monitoring is a technical problem to be solved at present. SUMMARY

[0005] The purpose of the present application is to solve the problem of low precision and adaptability of foundation pit deformation monitoring in the prior art, and a foundation pit deformation monitoring method applying image analysis is proposed, which comprises,

[0006] comprehensively dividing the engineering construction stage to generate a plurality of sequentially adjacent engineering construction stage processes, and determining the engineering construction stage related to the foundation pit on the engineering construction stage process;

[0007] confirming all deformation contents involved in the engineering construction stage related to the foundation pit, and dividing all deformation contents into direct deformation contents and indirect deformation contents according to the deformation subject;

[0008] setting up image monitoring points on the engineering construction stage related to the foundation pit, collecting deformation images of the engineering construction stage related to the foundation pit by means of the image monitoring points, identifying the deformation features of the direct deformation contents and the indirect deformation contents according to the deformation images, and determining the deformation degree of each engineering construction stage related to the foundation pit;

[0009] The deformation degree correlation between multiple foundation pit related engineering construction stages is analyzed, and a threshold of the deformation degree of each foundation pit related engineering construction stage is set, so as to realize the foundation pit deformation monitoring and early warning.

[0010] In some embodiments of the present application, all deformation contents are divided into direct deformation contents and indirect deformation contents according to deformation subjects, including,

[0011] All deformation subjects involved in each engineering construction stage are identified, including the foundation pit, the supporting structure, the surrounding building and the underground pipeline.

[0012] The deformation contents are divided into direct deformation contents and indirect deformation contents by the differences between the deformation subjects of the foundation pit, the supporting structure, the surrounding building and the underground pipeline.

[0013] In some embodiments of the present application, the deformation features of the direct deformation contents and the indirect deformation contents are identified according to the deformation images, including,

[0014] According to the engineering construction stage, the time point and the deformation subject to which the deformation image belongs, a deformation image time sequence of each deformation subject under each engineering construction stage is constructed, a feature template of each deformation subject is pre-configured, stable features of each image in the deformation image time sequence are extracted, and the stable features are matched with the feature template of the deformation subject to identify the deformation subject on each image.

[0015] On the basis of the deformation subject on each image, the contour line of the deformation subject is detected by an edge detection algorithm, and the deformation features of the direct deformation contents and the indirect deformation contents of the deformation subject are determined according to the deformation image time sequence and the contour line of the deformation subject on each image.

[0016] The stable features include one or more of SIFT, SURF, ORB, BRIEF and HOG features.

[0017] In some embodiments of the present application, the feature template of each deformation subject is pre-configured, including,

[0018] Each deformation subject contains multiple deformation parts, historical images of each deformation part are collected, and stable feature points are extracted on the historical images to mark the deformation part regions of the deformation subject on the historical images.

[0019] The stable feature points located in the deformation part regions are recorded as target stable feature points, and the stable feature points located outside the deformation part regions are recorded as non-target stable feature points.

[0020] Feature point matching is performed on the target stable feature points on multiple historical images, a transformation matrix of the registered images is calculated, a re-projection error is calculated, and a stability index is obtained.

[0021] obtaining a difference index by comparing the difference between the target stable feature points and the non-target stable feature points on the plurality of historical images;

[0022] calculating a pixel gradient variation degree of the target stable feature point according to a distance between the target stable feature point and an edge of the deformation part region, dividing the deformation part region according to a texture complexity, assigning a texture complexity degree of each target stable feature point, and determining a saliency index combining the pixel gradient variation degree and the texture complexity degree of the target stable feature point;

[0023] calculating a comprehensive screening index of the target stable feature point combining the stability index, the difference index and the saliency index of the target stable feature point;

[0024] ;

[0025] wherein, the comprehensive screening index of the i-th target stable feature point, 、 、 、 the stability index, the difference index and the saliency index of the i-th target stable feature point, 、 、 the maximum value among the three, is a first constant.

[0026] screening the target stable feature points in the deformation part region by the comprehensive screening index of the target stable feature points, and describing all deformation part conditions of the deformation subject by the screened target stable feature points, so as to configure a feature template of all deformation parts of each deformation subject.

[0027] In some embodiments of the present application, the contour line of the deformation subject is detected by an edge detection algorithm, including,

[0028] the edge detection algorithm includes a high threshold and a low threshold of image gradient intensity;

[0029] analyzing a high percentile of a gradient intensity histogram of the deformation image, and determining an initial value of each of the high threshold and the low threshold according to the high percentile;

[0030] determining a signal-to-noise ratio, a contrast and a brightness on the deformation image, determining an adjustment factor of each of the high threshold and the low threshold respectively by comprehensively considering the signal-to-noise ratio, the contrast and the brightness, adjusting the initial value of each of the high threshold and the low threshold based on the adjustment factor, and thus detecting the contour line of the deformation subject. ​​​

[0031] In some embodiments of the present application, the deformation features of the direct deformation content and the indirect deformation content of the deformation subject are determined according to the time sequence of the deformation images and the contour line of the deformation subject on each frame image, including,

[0032] The contour line of the deformation subject on each frame image is connected by a plurality of edge points;

[0033] The time interval is determined according to the properties of the deformation subject, and the time continuity of a plurality of frame images in the time sequence of the deformation images is evaluated according to the time interval, and the similarity between the plurality of frame images is calculated by the change of the contour line;

[0034] If the similarity is lower than the standard threshold, the difference between the edge points of the plurality of frame images is compared to determine the deformation features of the direct deformation content and the indirect deformation content of the deformation subject;

[0035] Otherwise, the difference between the contour lines of the plurality of frame images is compared to determine the deformation features of the direct deformation content and the indirect deformation content of the deformation subject.

[0036] In some embodiments of the present application, the deformation degree correlation between a plurality of foundation pit related construction stages is analyzed, including,

[0037] The deformation degree of the foundation pit related construction stage is divided into the deformation degree of a plurality of deformation subjects;

[0038] The finite element software is used to establish a foundation pit model of the construction stage, the position, shape and deformation degree of each deformation subject are marked in the foundation pit model, the material properties and boundary conditions in the foundation pit model are defined, and a plurality of foundation pit related construction stages are simulated on the foundation pit model, the deformation degree correlation between different deformation subjects under a single foundation pit related construction stage is determined, and the deformation degree correlation between a plurality of foundation pit related construction stages is determined.

[0039] In some embodiments of the present application, the threshold of the deformation degree of each foundation pit related construction stage is set, including,

[0040] A plurality of foundation pit deformation states under each foundation pit related construction stage are defined by Markov chain, and the state transition probability from the current foundation pit state to the foundation pit deformation state is calculated;

[0041] The deformation degree correlation between a plurality of foundation pit related construction stages is quantified, and the correlation strength of the deformation degree is determined;

[0042] The threshold of the deformation degree of each foundation pit related construction stage is set according to the state transition probability and the correlation strength of the deformation degree;

[0043] ;

[0044] wherein, is the threshold of the deformation degree of the first construction stage related to the foundation pit, is the sum of the deformation degrees of all the construction stages related to the foundation pit before the first construction stage, is the correlation strength of the deformation degree of the first construction stage related to the foundation pit to all the construction stages related to the foundation pit before the first construction stage, is the initial threshold of the deformation degree, represents the initial threshold of the deformation degree obtained by mapping the deformation influence situation of all the construction stages related to the foundation pit before the first construction stage, is the state transition probability of the first construction stage related to the foundation pit, is the second constant.

[0045] Correspondingly, the application also provides a foundation pit deformation monitoring system using image analysis, comprising,

[0046] a first module for comprehensively dividing the construction stages to generate a plurality of sequentially adjacent construction stage processes, and determining the construction stages related to the foundation pit on the construction stage processes;

[0047] a second module for confirming all the deformation contents involved in the construction stages related to the foundation pit, and dividing all the deformation contents into direct deformation contents and indirect deformation contents according to the deformation subjects;

[0048] a third module for setting image monitoring points on the construction stages related to the foundation pit, collecting deformation images of the construction stages related to the foundation pit by means of the image monitoring points, identifying the deformation features of the direct deformation contents and the indirect deformation contents according to the deformation images, and determining the deformation degrees of each of the construction stages related to the foundation pit;

[0049] a fourth module for analyzing the deformation degree correlation relationship between the plurality of construction stages related to the foundation pit, and setting the threshold of the deformation degree of each of the construction stages related to the foundation pit, so as to realize the foundation pit deformation monitoring and early warning.

[0050] Compared with the prior art, the application has the following beneficial effects:

[0051] 1、According to the deformation main body, all deformation contents are divided into direct deformation contents and indirect deformation contents, so as to analyze the deformation correlation between different deformation main bodies, and provide a reliable basis for the deformation condition of the subsequent engineering construction stage. According to the deformation characteristics of the direct deformation contents and the indirect deformation contents identified from the deformation image, the deformation characteristics are identified according to the time scale of the deformation image, and the deformation degree of the foundation pit related engineering construction stage is determined by combining the direct deformation contents and the indirect deformation contents, so as to comprehensively describe all deformation conditions of the foundation pit in the engineering construction stage.

[0052] 2、The deformation degree correlation between multiple foundation pit related engineering construction stages is analyzed, the deformation degree correlation between the engineering construction stages is determined according to the deformation correlation between all deformation main bodies in the engineering construction stage, so as to consider the deformation influence of the previous engineering construction stage on the subsequent engineering construction stage, set the threshold value of the deformation degree, improve the accuracy and adaptability of the foundation pit deformation monitoring, and ensure the subsequent management and safety prevention of the foundation pit engineering. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 A flowchart of a foundation pit deformation monitoring method using image analysis is provided for the present application;

[0054] Figure 2 A structural diagram of a foundation pit deformation monitoring system using image analysis is provided for the present application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments.

[0056] Referring to Figure 1 A foundation pit deformation monitoring method using image analysis includes the following steps,

[0057] Step S101, the engineering construction stage is fully divided to generate multiple sequentially adjacent engineering construction stage processes, and the foundation pit related engineering construction stage is determined on the engineering construction stage process.

[0058] In this embodiment, the monitoring of the deformation of the foundation pit usually runs through multiple stages (foundation pit related engineering construction stages) of the foundation pit engineering construction, but mainly concentrates on the following key periods:

[0059] Foundation pit excavation stage:

[0060] This is the core stage of foundation pit deformation monitoring. As the excavation depth increases, both the supporting structure and the surrounding soil will deform, so close monitoring is required. Generally, monitoring is required 1-2 times per day during earthwork excavation to ensure timely detection and handling of potential deformation risks.

[0061] Foundation pit supporting structure construction phase:

[0062] During the construction of the supporting structure, deformation monitoring is also required. This is mainly because the stability and strength of the supporting structure are directly related to the safety of the foundation pit. By monitoring the deformation of the supporting structure, the construction plan can be adjusted in a timely manner to ensure the stability and safety of the supporting structure.

[0063] Foundation pit backfilling phase:

[0064] After the foundation pit is backfilled, although most of the construction work has been completed, monitoring work still needs to continue. This is because after the foundation pit is backfilled, the soil may undergo consolidation settlement and other deformations, which may affect the surrounding environment and buildings. Therefore, deformation monitoring is also required during the foundation pit backfilling phase until the deformation tends to be stable.

[0065] Post-completion operation phase:

[0066] In some cases, foundation pit deformation monitoring may continue to the post-completion operation phase. This is mainly to assess the long-term impact of the foundation pit project on the surrounding environment and buildings, ensuring the safety and stability of the project.

[0067] It can be understood that the foundation pit-related conditions in the previous engineering construction phase will have an impact on the foundation pit deformation in the subsequent engineering construction phase, and this impact needs to be considered.

[0068] Step S102, confirm all deformation contents involved in the engineering construction phase related to the foundation pit, and divide all deformation contents into direct deformation contents and indirect deformation contents according to the deformation subject.

[0069] In some embodiments of the present application, dividing all deformation contents into direct deformation contents and indirect deformation contents according to the deformation subject includes,

[0070] Confirming all deformation subjects involved in each engineering construction phase, the deformation subject including the foundation pit, the supporting structure, the surrounding buildings and the underground pipelines;

[0071] Dividing the deformation contents into direct deformation contents and indirect deformation contents by the different deformation subjects of the foundation pit, the supporting structure, the surrounding buildings and the underground pipelines.

[0072] In this embodiment, the deformation bodies include foundation pits, supporting structures, surrounding buildings, underground pipelines, and the like, and the deformation of these bodies can reflect the deformation of the foundation pit. The direct deformation content is directly related to the deformation of the foundation pit, such as the settlement of the bottom of the foundation pit and the displacement of the slope of the foundation pit, the deformation of the supporting structure, and the like. The indirect deformation content is indirectly related to the deformation of the foundation pit, such as the settlement and inclination of surrounding buildings, the displacement and deformation of underground pipelines, and the like.

[0073] In step S103, image monitoring points are set on the construction stages of the foundation pit related project, deformation images of the construction stages of the foundation pit related project are collected by means of the image monitoring points, deformation features of the direct deformation content and the indirect deformation content are identified according to the deformation images, and the deformation degree of each construction stage of the foundation pit related project is determined.

[0074] In this embodiment, in order to effectively monitor the deformation of the foundation pit, image monitoring points need to be set on the construction stages of the foundation pit related project. These monitoring points should be arranged at key positions of the slope, the bottom, the supporting structure, and the surrounding buildings and underground pipelines of the foundation pit.

[0075] It can be understood that when the image monitoring points are set, only one deformation body is often contained in one image. In order to ensure the accuracy of subsequent image analysis, the deformation features, i.e., the specific deformation values such as the settlement and displacement, focus on the degree of change.

[0076] In some embodiments of the present application, the deformation features of the direct deformation content and the indirect deformation content are identified according to the deformation images, which include,

[0077] According to the construction stage, the time point, and the deformation body to which the deformation image belongs, a deformation image time sequence of each deformation body in each construction stage is constructed, a feature template of each deformation body is pre-configured, stable features of each image in the deformation image time sequence are extracted, and the stable features are matched with the feature template of the deformation body to identify the deformation body on each image;

[0078] On the basis of the deformation body on each image, the contour line of the deformation body is detected by means of an edge detection algorithm, and the deformation features of the direct deformation content and the indirect deformation content of the deformation body are determined according to the deformation image time sequence and the contour line of the deformation body on each image.

[0079] The stable features include one or more of SIFT, SURF, ORB, BRIEF, and HOG features.

[0080] In this embodiment, the feature template of each deformation body is pre-configured, and reliable stable feature points on historical images are screened to serve as the basis of the template.

[0081] In some embodiments of the present application, a feature template of each deformable subject is pre-configured, including:

[0082] Each deformed subject contains multiple deformed parts. Historical images of each deformed part are collected, and stable feature points are extracted from the historical images to mark the deformed part area of ​​the deformed subject on the historical images.

[0083] The stable feature points located within the deformation area are recorded as target stable feature points, and the stable feature points located outside the deformation area are recorded as non-target stable feature points;

[0084] By matching the target stable feature points on multiple historical images, the transformation matrix of the registered image is calculated, and the reprojection error is calculated to obtain the stability index;

[0085] Compare the differences between target stable feature points and non-target stable feature points on multiple historical images to obtain a difference index;

[0086] The pixel gradient change degree of the target stable feature point is calculated based on the distance between the target stable feature point and the edge of the deformed area. The deformed area is divided according to the texture complexity. The texture complexity degree of each target stable feature point is assigned. The significant index is determined by combining the pixel gradient change degree and texture complexity of the target stable feature point.

[0087] The comprehensive screening index of the target stable feature point is calculated by combining the stability index, difference index and significance index of the target stable feature point;

[0088] ;

[0089] in, For the Comprehensive screening index of target stable feature points, 、 、 are the combined weights of stability index, difference index and significance index respectively, 、 、 For the The stability index, difference index and significance index of the target stable feature point, for The maximum of the three, is the first constant;

[0090] The target stable feature points within the deformation area are screened by using the comprehensive screening index of the target stable feature points. The screened target stable feature points are used to describe the conditions of all deformation parts of the deformation subject, thereby configuring the feature templates of all deformation parts of each deformation subject.

[0091] In this embodiment, each deformed body contains multiple deformed parts, for example, a foundation pit itself can include pit bottom, pit body and pit top and other parts. For multiple parts of each deformed body object, feature descriptors are defined according to the features extracted in its historical images. The template can be a set of feature points or a set of feature descriptors, which is used to represent the feature information of the part. The defined template is stored for subsequent matching with the features in the image to be recognized. The storage mode can be file, database, etc., depending on the application scenario and requirements.

[0092] In this embodiment, the stability index, difference index and saliency index of the stable feature points respectively describe the stability (select those feature points that can remain stable under different image transformations such as rotation, scaling, illumination change), uniqueness (preferentially select those feature points with significant uniqueness, which can be easily distinguished in the image.) and saliency (the saliency of a feature point can be measured by the gradient change or texture complexity of its surrounding pixels, and the feature point with high saliency is more likely to play a key role in image matching) of the feature points, which are obviously beneficial to the recognition of the deformed body.

[0093] In this embodiment, feature point matching is performed using feature descriptors, which can use KNN (nearest neighbor) algorithm or FLANN (fast library for nearest neighbor search) for matching. The matched feature points are used to register the images, and RANSAC algorithm can be used to estimate the transformation model and eliminate the wrong matching feature points. The transformation matrix of the registered image is calculated, and the stability of the feature points is evaluated by calculating the re-projection error. The smaller the re-projection error, the more stable the feature points.

[0094] In this embodiment, the Euclidean distance or normalized cross-correlation between the feature descriptors (the difference between the target stable feature points and the non-target stable feature points) is calculated to measure the difference between the feature points. The greater the distance or the smaller the normalized cross-correlation, the more unique the feature points.

[0095] In this embodiment, the gradient change of the pixels around the feature point is calculated, and the feature point with greater gradient change is more salient. The texture complexity of the region where the feature point is located is analyzed, and the region with more complex texture usually contains more information, so the feature point is more salient. Because texture is a feature that multiple pixels jointly represent, the deformed part region is divided according to the texture complexity, and the texture complexity of each stable feature point is assigned, and the saliency index is determined by combining the pixel gradient change degree and texture complexity (weighted sum, etc.) of the target stable feature point.

[0096] In this embodiment, represents the correction of the maximum value in the stability index, difference index and saliency index to the average value of the three, is to balance the size of the correction function.

[0097] In some embodiments of the present application, the contour line of the deformed object is detected by an edge detection algorithm, including,

[0098] The edge detection algorithm includes a high threshold and a low threshold of image gradient intensity;

[0099] The high percentile of the gradient intensity histogram of the deformed image is analyzed, and the initial values of the high threshold and the low threshold are determined according to the high percentile;

[0100] The signal-to-noise ratio, contrast and brightness on the deformed image are determined, the adjustment factors of the high threshold and the low threshold are respectively determined by comprehensively considering the signal-to-noise ratio, contrast and brightness, and the initial values of the high threshold and the low threshold are adjusted based on the adjustment factors, so as to detect the contour line of the deformed object.

[0101] In the present embodiment, the contour line, also known as the external line, is a line describing the external edge of an object. In image processing and computer vision, contour lines are often used to represent the shape and boundary of an object (part of the deformed object). The contour line is composed of edge points. The edge point is a pixel point in the image where the gray value changes sharply. These points are connected to form a contour line.

[0102] In the present embodiment, the threshold refers to the threshold of image gradient intensity. The Canny edge detection algorithm distinguishes between strong edges, weak edges and noise edges by setting two thresholds (threshold1 and threshold2). Edge detection algorithms, such as Sobel operator, Prewitt operator and Canny operator, can detect edge points in an image. These algorithms are based on the difference between the local region and the surrounding region of the image to find the position where the gray value changes significantly. In complex images, multiple interference contours may be extracted, increasing the difficulty of subsequent processing. The solution is to set appropriate thresholds to filter and remove interference contours. The high threshold (threshold2) is used to detect strong edges. If it is set too high, some important edges may be missed; if it is set too low, too much noise may be detected. Generally, the high threshold can be set to the high percentile of the image gradient intensity, such as 90% or 95%. The low threshold (threshold1) is used to connect edges. If it is set too high, weak edges may be missed; if it is set too low, too much noise may be introduced. The common practice is to set the low threshold to a certain proportion of the high threshold, usually between 1:2 and 1:3.

[0103] In this embodiment, the distribution of gradient intensity can be understood by calculating the gradient intensity histogram of the image. The high threshold can be usually set as a high percentile (such as 90% or 95%) of the gradient intensity histogram to ensure that only strong edges are detected. The low threshold can be set as a certain proportion (such as between 1:2 and 1:3) of the high threshold to connect weak edges. In images with high contrast and brightness, edges can be more obvious, so the threshold can be appropriately increased. In images with low contrast and brightness, edges can be more blurred, so the threshold needs to be reduced. Similarly, the influence of signal-to-noise ratio on edges, images with more noise need to set a lower threshold to avoid misjudging noise as edges. Conversely, images with less noise can set a higher threshold to improve the accuracy of edge detection. According to the corresponding adjustment factors of the high threshold and the low threshold respectively determined by the signal-to-noise ratio, the contrast and the brightness, the adjustment is completed by adjusting the initial value by the adjustment factor.

[0104] In some embodiments of the present application, the deformation characteristics of the direct deformation content and the indirect deformation content of the deformation subject are determined according to the time sequence of the deformation image and the contour line of the deformation subject on each frame of image, including,

[0105] The contour line of the deformation subject on each frame of image is connected by a plurality of edge points;

[0106] The time interval is determined according to the nature of the deformation subject, and the time continuity of the plurality of frames of image on the time sequence of the deformation image is evaluated, and the similarity between the plurality of frames of image is calculated by the change of the contour line;

[0107] If the similarity is lower than the standard threshold, the difference between the edge points of the plurality of frames of image is compared to determine the deformation characteristics of the direct deformation content and the indirect deformation content of the deformation subject;

[0108] Otherwise, the difference between the contour lines of the plurality of frames of image is compared to determine the deformation characteristics of the direct deformation content and the indirect deformation content of the deformation subject.

[0109] In this embodiment, the time interval is determined according to the nature of the deformation subject (referring to the stable condition), when the deformation change degree of the deformation subject between the plurality of frames of image is large, the change of the contour line can clearly describe the deformation, when the deformation change degree is small, the change of the contour line is not clear, at this time the change of the edge points is used to describe the deformation. Whether the change is obvious is judged by calculating the shape similarity of the object contour. If the shape similarity is low, it means that the shape of the object has changed significantly. When the local detail change of the object edge needs to be paid attention to, such as crack propagation and small deformation, the change of the edge points should be paid attention to.

[0110] In this embodiment, the deformation characteristics include shape characteristics, position characteristics, gray characteristics, time attributes, etc.

[0111] Shape features: including the length, area, perimeter, curvature of the contour line, etc., which can describe the overall shape and edge details of the object.

[0112] Position features: the position changes of the object in the image, such as translation, rotation, scaling, etc., which can reflect the motion state of the object.

[0113] Gray features: the gray value changes of the edge points, which can reflect the brightness or color changes of the object edge.

[0114] Time features: the trend of deformation features over time, such as deformation rate, deformation acceleration, etc., which can reveal the dynamic process of object deformation.

[0115] It should be noted that to determine the deformation degree of each foundation pit related construction stage, first determine the deformation features of all deformation subjects under the construction stage, combine the deformation features according to direct deformation content and indirect deformation content, and determine the deformation degree of the construction stage.

[0116] Step S104, analyze the deformation degree correlation between multiple foundation pit related construction stages, set the threshold value of the deformation degree of each foundation pit related construction stage, and realize the foundation pit deformation monitoring and early warning.

[0117] In this embodiment, the simulation software is used to simulate the multiple stages and multiple deformation subjects of the foundation pit, first determine the correlation between the multiple deformation subjects, and then determine the deformation degree correlation between the multiple construction stages according to the correlation between the multiple deformation subjects.

[0118] In some embodiments of the present application, analyzing the deformation degree correlation between multiple foundation pit related construction stages includes,

[0119] Splitting the deformation degree of the foundation pit related construction stage into the deformation degree of multiple deformation subjects;

[0120] Using finite element software to establish a foundation pit model of the construction stage, marking the position, shape and deformation degree of each deformation subject in the foundation pit model, defining the material properties and boundary conditions in the foundation pit model, and simulating multiple foundation pit related construction stages on the foundation pit model, determining the deformation degree correlation between different deformation subjects under a single foundation pit related construction stage, and then determining the deformation degree correlation between multiple foundation pit related construction stages.

[0121] In this embodiment, through simulation, the deformation degree correlation between different deformation subjects is identified. The deformation subjects will affect each other, and then the deformation degree correlation between multiple foundation pit-related engineering construction stages is determined, and the impact of the deformation situation in the previous stage on the deformation situation in the subsequent stage is identified.

[0122] In some embodiments of the present application, the threshold value of the degree of deformation of each foundation pit-related engineering construction stage is set, including:

[0123] The Markov chain is used to define several pit deformation states under each pit-related construction stage, and the state transition probability between the current pit state and the pit deformation state is calculated;

[0124] Quantify the correlation between the deformation degrees of multiple foundation pit-related construction stages and determine the correlation strength of the deformation degrees;

[0125] The threshold value of the deformation degree of each foundation pit-related engineering construction stage is set according to the correlation strength between the state transition probability and the deformation degree;

[0126] ;

[0127] in, For the The threshold value of deformation degree during the construction phase of the foundation pit-related engineering, For the The sum of the deformation degrees of all previous foundation pit-related construction stages, For the All previous foundation pit related construction phases are for The correlation strength of the deformation degree of each foundation pit related engineering construction stage, is the initial threshold of deformation degree, Indicates that The initial threshold value of deformation degree is obtained by mapping the deformation influence of all previous foundation pit related engineering construction stages. For the The state transition probability of the engineering construction phase related to the foundation pit, is the second constant.

[0128] In this embodiment, a Markov chain is a stochastic process model that assumes that the probability of a state transition at a given moment depends solely on the previous state. In foundation pit deformation analysis, Markov chains can be used to describe the dynamic evolution of the foundation pit deformation system. This is particularly useful when the state transition probability of the deformation system has no aftereffects. This allows for calculating the probability of transitioning from the current state of the current construction phase to a future deformation state.

[0129] In this embodiment, represents the influence amount of the previous engineering construction stage on the current stage, maps an initial threshold value of a deformation degree, and the greater the influence amount, the smaller the initial threshold value, represents the correction of the initial threshold value by the state transition probability, is the maximum one of the plurality of foundation pit deformation state probability values, is determined by the foundation pit deformation state and exists to balance the size of the correction function.

[0130] Correspondingly, the application also provides a foundation pit deformation monitoring system applying image analysis, as shown in Figure 2 , comprising,

[0131] The first module is used for comprehensively dividing the engineering construction stages to generate a plurality of sequentially adjacent engineering construction stage processes, and determining the engineering construction stages related to the foundation pit on the engineering construction stage processes.

[0132] The second module is used for confirming all deformation contents involved in the engineering construction stages related to the foundation pit, and dividing all the deformation contents into direct deformation contents and indirect deformation contents according to deformation subjects.

[0133] The third module is used for setting image monitoring points on the engineering construction stages related to the foundation pit, collecting deformation images of the engineering construction stages related to the foundation pit by means of the image monitoring points, identifying deformation features of the direct deformation contents and the indirect deformation contents according to the deformation images, and determining deformation degrees of each engineering construction stage related to the foundation pit.

[0134] The fourth module is used for analyzing deformation degree correlation relationships between the plurality of engineering construction stages related to the foundation pit, setting threshold values of the deformation degrees of each engineering construction stage related to the foundation pit, so as to realize foundation pit deformation monitoring and early warning.

[0135] Compared with the prior art, the application has the following beneficial effects:

[0136] 1. According to the application, all deformation contents are divided into direct deformation contents and indirect deformation contents according to deformation subjects, so as to analyze deformation correlations between different deformation subjects, and provide a reliable basis for deformation conditions of subsequent engineering construction stages. According to the application, deformation features of the direct deformation contents and the indirect deformation contents are identified according to time scales of deformation images, and the deformation degrees of the engineering construction stages related to the foundation pit are determined by combining the direct deformation contents and the indirect deformation contents, so that all deformation conditions of the engineering construction stages related to the foundation pit are comprehensively described.

[0137] 2、Analyzing the deformation degree correlation between multiple foundation pit related engineering construction stages, determining the deformation degree correlation between engineering construction stages by the deformation correlation between all deformation subjects in the engineering construction stage, thereby considering the deformation influence of the previous engineering construction stage on the subsequent engineering construction stage, setting the threshold of deformation degree, improving the accuracy and adaptability of foundation pit deformation monitoring, and ensuring the subsequent management and safety prevention of foundation pit engineering.

[0138] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by hardware, or by means of software and necessary general hardware platform. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.), and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present application.

[0139] Those skilled in the art can understand that the drawings are only a schematic diagram of a preferred embodiment, and the modules or processes in the drawings are not necessarily required for implementing the present application.

[0140] Those skilled in the art can understand that the modules in the system in the embodiments can be distributed in the system in the embodiments according to the description of the embodiments, or can be changed and located in one or more systems different from the embodiments. The modules in the above embodiments can be combined into one module, or can be further split into multiple sub-modules.

[0141] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can make equivalent replacements or changes to the technical solutions and inventive concepts of the present application within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for monitoring deformation of a foundation pit using image analysis, characterized by, The method comprises the following steps: comprehensively dividing the engineering construction stage to generate a plurality of sequentially adjacent engineering construction stage processes, and determining the engineering construction stage related to the foundation pit on the engineering construction stage process; confirming all deformation contents involved in the engineering construction stage related to the foundation pit, and dividing all deformation contents into direct deformation contents and indirect deformation contents according to deformation subjects; setting up image monitoring points on the engineering construction stage related to the foundation pit, collecting deformation images of the engineering construction stage related to the foundation pit by means of the image monitoring points, identifying the deformation features of the direct deformation contents and the indirect deformation contents according to the deformation images, and determining the deformation degree of each engineering construction stage related to the foundation pit; analyzing the deformation degree correlation between a plurality of engineering construction stages related to the foundation pit, and setting the threshold value of the deformation degree of each engineering construction stage related to the foundation pit, so as to realize the foundation pit deformation monitoring and early warning; identifying the deformation features of the direct deformation contents and the indirect deformation contents according to the deformation images, which comprises constructing the deformation image time sequence of each deformation subject in each engineering construction stage according to the engineering construction stage, time point and deformation subject to which the deformation image belongs, pre-configuring the feature template of each deformation subject, extracting stable features from each image in the deformation image time sequence, and matching the stable features with the feature template of the deformation subject to identify the deformation subject on each image; on the basis of the deformation subject on each image, detecting the contour line of the deformation subject by means of an edge detection algorithm, and determining the deformation features of the direct deformation contents and the indirect deformation contents of the deformation subject according to the deformation image time sequence and the contour line of the deformation subject on each image; wherein the stable features comprise one or more of SIFT, SURF, ORB, BRIEF and HOG features; the pre-configuration of the feature template of each deformation subject comprises each deformation subject comprises a plurality of deformation parts, historical images of each deformation part are collected, and stable feature points are extracted from the historical images to mark the deformation part area of the deformation subject on the historical images; the stable feature points located in the deformation part area are recorded as target stable feature points, and the stable feature points located outside the deformation part area are recorded as non-target stable feature points; feature point matching is performed on the target stable feature points on a plurality of historical images, the transformation matrix of the registered images is calculated, the re-projection error is calculated, and the stability index is obtained; the differences between the target stable feature points and the non-target stable feature points on a plurality of historical images are compared, and the difference index is obtained; the pixel gradient change degree of the target stable feature points is calculated according to the distance between the target stable feature points and the edge of the deformation part area, the deformation part area is divided according to the texture complexity, the texture complexity of each target stable feature point is allocated, and the saliency index is determined by combining the pixel gradient change degree and the texture complexity of the target stable feature points; The comprehensive screening index of the target stable feature point is calculated by combining the stability index, difference index and significance index of the target stable feature point; ;in, For the Comprehensive screening index of target stable feature points, are the combined weights of stability index, difference index and significance index respectively, For the The stability index, difference index and significance index of the target stable feature point, for The maximum of the three, is the first constant; the target stable feature points within the deformation area are screened by the comprehensive screening index of the target stable feature points, and the screened target stable feature points are used to describe the conditions of all deformed parts of the deformed subject, thereby configuring the feature templates of all deformed parts of each deformed subject; the analysis of the deformation degree correlation between a plurality of engineering construction stages related to the foundation pit comprises splitting the deformation degree of the engineering construction stage related to the foundation pit into the deformation degree of a plurality of deformation subjects. Use finite element software to establish a foundation pit model for the construction phase. Mark the location, shape, and deformation degree of each deformed entity in the foundation pit model. Define the material properties and boundary conditions in the foundation pit model. Simulate multiple foundation pit-related construction phases on the foundation pit model to determine the correlation between the deformation degrees of different deformed entities in a single foundation pit-related construction phase. Then, determine the correlation between the deformation degrees of multiple foundation pit-related construction phases. The threshold value for setting the deformation degree of each foundation pit-related engineering construction stage includes: The Markov chain is used to define several pit deformation states under each pit-related construction stage, and the state transition probability between the current pit state and the pit deformation state is calculated; Quantify the correlation between the deformation degrees of multiple foundation pit-related construction stages and determine the correlation strength of the deformation degrees; setting the threshold value of the deformation degree of each foundation-related engineering construction stage according to the state transition probability and the correlation intensity of the deformation degree; ; wherein, is the threshold value of the deformation degree of the first foundation-related engineering construction stage, is the sum of the deformation degrees of all the foundation-related engineering construction stages before the first foundation-related engineering construction stage, is the correlation intensity of the deformation degree of the first foundation-related engineering construction stage to all the foundation-related engineering construction stages before the first foundation-related engineering construction stage, is the initial threshold value of the deformation degree, represents the initial threshold value of the deformation degree obtained by mapping the deformation influence situation of all the foundation-related engineering construction stages before the first foundation-related engineering construction stage, is the state transition probability of the first foundation-related engineering construction stage, is the second constant.

2. The method for monitoring deformation of a foundation pit using image analysis according to claim 1, wherein, All deformation contents are divided into direct deformation contents and indirect deformation contents according to the deformation subject, including: Identify all deformation entities involved in each construction phase, including foundation pits, supporting structures, surrounding buildings, and underground pipelines; The deformation contents are divided into direct deformation contents and indirect deformation contents based on the differences in the deformation subjects of foundation pit, supporting structure, surrounding buildings and underground pipelines.

3. The method for monitoring foundation pit deformation using image analysis according to claim 1, characterized in that: The contour of the deformed subject is detected by edge detection algorithm, including, The edge detection algorithm includes high and low thresholds for the image gradient intensity; Analyzing a high percentile of a gradient intensity histogram of the deformed image, and determining initial values ​​of a high threshold and a low threshold according to the high percentile; The signal-to-noise ratio, contrast and brightness on the deformed image are determined, and the adjustment factors of the high threshold and the low threshold are determined respectively based on the comprehensive signal-to-noise ratio, contrast and brightness. The initial values ​​of the high threshold and the low threshold are adjusted based on the adjustment factors to detect the contour line of the deformed subject.

4. The method for foundation pit deformation monitoring using image analysis according to claim 1, wherein, Determine the deformation features of the direct deformation content and indirect deformation content of the deformation subject according to the deformation image time series and the contour line of the deformation subject on each frame image, including: The contour line of the deformed subject in each frame image is connected by multiple edge points; Determine the time interval according to the properties of the deformed subject, perform time-continuous evaluation on multiple frames of the deformed image time series according to the time interval, and calculate the similarity between the multiple frames of images through the change of the contour lines; If the similarity is lower than the standard threshold, the differences in edge points between multiple frames of images are compared to determine the deformation features of the direct deformation content and indirect deformation content of the deformation subject; Otherwise, the differences in contour lines between the multiple frames of images are compared to determine the deformation features of the direct deformation content and the indirect deformation content of the deformation subject.

5. A foundation pit deformation monitoring system using image analysis, characterized by, The method for monitoring foundation pit deformation according to claim 1 comprises: The first module is used to comprehensively divide the engineering construction phases, generate multiple sequentially adjacent engineering construction phase processes, and determine the engineering construction phases related to the foundation pit in the engineering construction phase processes; The second module is used to confirm all deformation contents involved in the construction phase of the foundation pit-related engineering, and divide all deformation contents into direct deformation contents and indirect deformation contents according to the deformation subject; The third module is configured to set image monitoring points on the construction stages of the foundation pit related projects, collect deformation images of the construction stages of the foundation pit related projects by means of the image monitoring points, identify deformation features of direct deformation contents and indirect deformation contents according to the deformation images, and determine deformation degrees of each construction stage of the foundation pit related projects; The fourth module is configured to analyze a deformation degree correlation relationship between the construction stages of the foundation pit related projects, set a threshold value of the deformation degree of each construction stage of the foundation pit related projects, and thus realize the foundation pit deformation monitoring and early warning.

Images