Slope displacement prediction method and device

Abstract

The embodiments of this specification provide a slope displacement prediction method and device, wherein the slope displacement prediction method includes: obtaining an initial graph structure of the monitored area before the current moment, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points; simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain an intermediate graph structure; using a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information, and using a recurrent neural network to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment; inputting the time information of the current moment into a regressor to obtain the displacement prediction result of the current moment. This method can not only efficiently process large-scale spatial data, but also accurately capture and predict complex dynamic relationships in time series data, providing strong technical support for the safety monitoring of open-pit coal mine slopes.

Description

Technical Field

[0001] The embodiments of this specification relate to the technical field of slope displacement prediction, and in particular to a slope displacement prediction method. Background Art

[0002] In recent years, the proportion of open-pit mining in coal mining has increased. The open-pit coal mine slopes formed during excavation are easily subject to the risk of excessive deformation and instability of open-pit coal mine slopes under the influence of internal and external forces such as the strong development of ultra-thick coal seam structural surfaces and strong freeze-thaw, posing a huge threat to the high-quality and safe mining of open-pit coal mines.

[0003] Current slope stability prediction methods are highly dependent on comprehensive and detailed data collection, aiming to improve the accuracy and reliability of slope safety assessment by integrating diversified information. This process not only involves traditional geological survey data, such as geotechnical physical and mechanical properties, geological structural characteristics, etc., but also includes advanced real-time monitoring data, such as GPS displacement monitoring, remote sensing image analysis, underground seepage and stress monitoring and other high-tech means. In addition, environmental factors are also taken into consideration, which may have a direct or indirect impact on slope stability. The integration and analysis of all these massive data constitute a complex data processing chain, which aims to identify the key factors affecting slope safety and their interaction mechanisms through deep mining of big data technology and machine learning algorithms, so as to achieve accurate prediction and early warning of the future state of the slope, and ensure the safety and efficiency of mining operations.

[0004] However, this slope stability prediction method, which is highly dependent on comprehensive and detailed data collection, brings a series of challenges, such as the inability to efficiently process large amounts of data and the inability to deeply capture the internal potential connections of large amounts of data. Therefore, a more efficient and practical slope displacement prediction method is urgently needed. Summary of the invention

[0005] In view of this, an embodiment of this specification provides a slope displacement prediction method. One or more embodiments of this specification also relate to a slope displacement prediction device, a computing device, a computer-readable storage medium and a computer program to solve the technical defects existing in the prior art.

[0006] According to a first aspect of an embodiment of this specification, a slope displacement prediction method is provided, comprising:

[0007] Acquire an initial graph structure of the area to be monitored before the current time, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points;

[0008] Simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure;

[0009] Analyzing the spatial features of the intermediate graph structure using a graph neural network to obtain spatial information, and analyzing the spatial information and time information at a previous moment using a recurrent neural network to obtain time information at a current moment;

[0010] The time information of the current moment is input into the regressor to obtain the displacement prediction result of the current moment.

[0011] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the step of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure includes:

[0012] Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0013] Acquire location information of the first monitoring point and the second monitoring point, and determine a distance between the first monitoring point and the second monitoring point according to the location information;

[0014] When the interval distance is less than a distance threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0015] When the interval distance is not less than a distance threshold, the edge between the first monitoring point and the second monitoring point is removed.

[0016] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the method of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure further includes:

[0017] Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0018] Obtaining the cumulative displacement of the first monitoring point and the second monitoring point within a preset time and the displacement at the current moment;

[0019] Determine the displacement connection strength between the first monitoring point and the second monitoring point according to the displacement accumulation amount and the displacement amount at the current moment;

[0020] When the displacement connection strength is less than a connection strength threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0021] When the displacement connection strength is not less than a connection strength threshold, the edge between the first monitoring point and the second monitoring point is removed.

[0022] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the method of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure further includes:

[0023] Determine the adjacency matrix between each monitoring point;

[0024] Obtaining a transfer matrix between each monitoring point based on the adjacency matrix;

[0025] Determine the displacement characteristic diffusion matrix between the monitoring points according to the transfer matrix and the spatial diffusion intensity;

[0026] Determine a first monitoring point and a second monitoring point, and determine target displacement feature diffusion values ​​corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0027] When the target displacement feature diffusion value is less than a preset spatial diffusion threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0028] When the target displacement feature diffusion value is not less than a preset spatial diffusion threshold, an edge between the first monitoring point and the second monitoring point is removed.

[0029] In one or more embodiments of the present specification, after simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure, the method further includes:

[0030] Obtaining a preset time lag range and a preset time lag intensity;

[0031] determining a diffusion intensity variation function according to the time lag range and the time lag intensity;

[0032] The intermediate graph structure data is subjected to time diffusion convolution according to the diffusion intensity variation function.

[0033] In one or more embodiments of the present specification, the obtaining of spatial information by analyzing the spatial features of the intermediate graph structure using a graph neural network includes:

[0034] Acquire the target monitoring point and a neighbor set of the target monitoring point;

[0035] The graph neural network is used to aggregate the spatial features of the target monitoring point and the neighbor set of the target monitoring point to obtain the spatial information of the target monitoring point at the current moment.

[0036] In one or more embodiments of the present specification, the step of analyzing the spatial information and the time information at the previous moment by using a recurrent neural network to obtain the time information at the current moment includes:

[0037] Obtaining the time information of the monitoring point at the previous moment;

[0038] The updating function of the recurrent neural network is used to update the spatial information of the target monitoring point at the current moment and the time information of the monitoring point at the previous moment to obtain the time information at the current moment.

[0039] According to a second aspect of an embodiment of this specification, a slope displacement prediction device is provided, comprising:

[0040] An acquisition module is configured to acquire an initial graph structure of the area to be monitored at a current moment, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points;

[0041] an intermediate graph structure obtaining module, configured to simplify the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure;

[0042] A time information acquisition module is configured to use a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information, and use a recurrent neural network to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment;

[0043] The displacement prediction result obtaining module is configured to input the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment.

[0044] According to a third aspect of an embodiment of this specification, a computing device is provided, including:

[0045] Memory and processor;

[0046] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the above slope displacement prediction method are implemented.

[0047] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned slope displacement prediction method are implemented.

[0048] According to a fifth aspect of the embodiments of this specification, a computer program product is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned slope displacement prediction method.

[0049] An embodiment of the present specification realizes the reduction of the connection density of the monitoring point graph by simplifying the connection relationship of the monitoring points, and uses the graph diffusion and time diffusion methods to perform diffusion modeling on the time-invariant characteristics and time-varying characteristics of the monitoring points, captures the time lag effect of the time-varying characteristics on the displacement, captures its spatiotemporal relationship through the spatiotemporal model, and obtains the displacement prediction result of the model. It can not only efficiently process large-scale spatial data, but also accurately capture and predict the complex dynamic relationships in time series data, providing strong technical support for the safety monitoring of open-pit coal mine slopes. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of a slope displacement prediction method provided by an embodiment of this specification;

[0051] Figure 2 It is a graph structure of a slope displacement prediction method provided by an embodiment of this specification;

[0052] Figure 3 It is a convolution process of a slope displacement prediction method provided by an embodiment of this specification;

[0053] Figure 4 It is a flow chart of a slope displacement prediction method processing process provided by an embodiment of this specification;

[0054] Figure 5 It is a structural schematic diagram of a slope displacement prediction device provided by an embodiment of this specification;

[0055] Figure 6 It is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION

[0056] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0057] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0058] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0059] In addition, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0060] First, the terms involved in one or more embodiments of this specification are explained.

[0061] Graph Neural Networks (GNNs) is a deep learning method based on graph structure.

[0062] Recurrent Neural Network: (Recurrent Neural Network) is a recurrent neural network used to process sequence data.

[0063] In this specification, a slope displacement prediction method is provided. This specification also relates to a slope displacement prediction device, a computing device, and a computer-readable storage medium, which are described in detail one by one in the following embodiments.

[0064] See also Figure 1 , Figure 1 A flow chart of a slope displacement prediction method provided according to an embodiment of the present specification is shown, which specifically includes steps 102 to 108.

[0065] Step 102: Acquire an initial graph structure of the area to be monitored before the current time, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points.

[0066] Specifically, the areas to be monitored refer to any hillside or slope areas that require safety assessment, stability monitoring and maintenance. Such areas may be formed by natural factors (such as natural slopes formed by mountain and river erosion) or human factors (such as road cuttings, embankments, mining slopes and other artificial slopes formed by excavation and filling).

[0067] In specific implementation, monitoring can be carried out through a variety of monitoring means, such as: monitoring the stress distribution and changes inside the rock and soil body through instruments such as strain gauges and stress gauges, and using modern technologies such as remote sensing technology, laser scanning (LiDAR), radar and drone photogrammetry to quickly monitor terrain changes over a large range. Use rainfall gauges, thermometers, groundwater level gauges, etc. to monitor external environmental conditions that affect the stability of the slope. Use seismometers and accelerometers to record the vibration effects caused by seismic activities or artificial blasting. In the process of monitoring through a variety of monitoring means, it is necessary to set up multiple monitoring points. Monitoring points refer to fixed positions set at specific locations on the slope or its surroundings for installing various monitoring instruments and / or markers. The setting of monitoring points needs to fully consider factors such as the geological conditions, slope and slope height of the slope, and be determined in combination with the purpose and requirements of monitoring. When setting monitoring points, the layout of monitoring instruments should be taken into account, and the monitoring instruments need to choose appropriate locations to ensure the accuracy of monitoring data as much as possible. In specific implementation, those skilled in the art can determine according to their own monitoring purposes and requirements.

[0068] Multiple monitoring points are set up in a monitored area. These points are carefully selected and arranged to accurately and continuously collect information related to the stability of the slope.

[0069] After obtaining the original data collected by various monitoring means at each monitoring point, the attribute information corresponding to each monitoring point is determined respectively, and the connection relationship between each monitoring point is established; after determining the above information, an initial graph structure is established according to the attribute information corresponding to the monitoring point and the connection relationship between each monitoring point. The initial graph structure includes nodes and edges, wherein a monitoring point is a node, and an edge refers to the connection relationship between monitoring points.

[0070] In particular, it can be assumed that any two monitoring points in the monitored area have a connection relationship. For example, the monitored area includes monitoring point A, monitoring point B, monitoring point C, and monitoring point D. Then, when establishing the initial graph structure, a certain monitoring point is associated with the other three monitoring points. Taking monitoring point D as an example, it is associated with monitoring point A, monitoring point B, and monitoring point C, and finally the following is obtained: Figure 2 The initial graph structure shown.

[0071] Specifically, the initial graph structure may include multiple ones. In the process of monitoring the slope, data collected at each monitoring point will be obtained at preset time intervals. Each time data is collected at a moment, a corresponding initial graph structure will be generated. The initial graph structure contains the attribute information corresponding to the monitoring point at the current moment. In practical applications, those skilled in the art can determine according to their own needs, such as: the preset time interval is 1 minute, the preset time period is 20 minutes, then data will be collected at each monitoring point every minute, and the slope displacement will be predicted based on the data collected 20 minutes before the current moment. There will be 20 initial graph structures, and the slope displacement will be predicted based on these 20 initial graph structures. Of course, the preset time interval and the preset time period can be of any length, and the preset time interval and the preset time period can be 1 hour, 24 hours, 1 week, etc., and this manual does not limit this.

[0072] In one or more embodiments of the present specification, the initial graph structure includes at least two monitoring points and attribute information corresponding to each monitoring point, wherein the attribute information includes at least time-invariant features of the monitoring points, and may also include time-varying features, etc. The time-invariant features may be the location information of the monitoring points, or the displacement of the monitoring points at the current moment relative to the previous moment.

[0073] Based on this, after obtaining the original slope data collected by each monitoring point in the monitored area, an initial graph structure is established according to the original data. The initial graph structure includes multiple monitoring points and the connection relationship between the monitoring points, the location information of each monitoring point, and the displacement of the monitoring point at the current moment relative to the previous moment. When performing slope displacement prediction, the initial graph structure is obtained.

[0074] Step 104: Simplify the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure.

[0075] The initial graph structure of the area to be monitored is obtained, and the connection relationship between the monitoring points is simplified by using the time-invariant characteristics of the monitoring points in the initial graph structure to obtain the intermediate graph structure.

[0076] Specifically, the intermediate graph structure refers to the graph structure after simplifying the connection relationship step, such as Figure 2 The intermediate graph structure shown simplifies the connection relationship between monitoring points based on the time-invariant characteristics. By analyzing the inherent time-invariant characteristics of the monitoring points, the connection relationship between the monitoring points in the initial graph structure is optimized and simplified, and then a more refined intermediate graph structure is extracted. This process not only deepens the understanding of the stable association between monitoring points, but also promotes the efficient expression of the graph structure, providing a clearer and more concentrated information view for subsequent analysis and decision-making.

[0077] Considering that two monitoring points that are far apart have limited influence on each other, whether the monitoring points have a connection relationship can be determined according to the distance between the monitoring points, and a distance threshold is set. If the distance threshold is exceeded, the connection relationship is discarded, otherwise it is retained.

[0078] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the step of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure includes:

[0079] Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0080] Acquire location information of the first monitoring point and the second monitoring point, and determine a distance between the first monitoring point and the second monitoring point according to the location information;

[0081] When the interval distance is less than a distance threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0082] When the interval distance is not less than a distance threshold, the edge between the first monitoring point and the second monitoring point is removed.

[0083] Specifically, the first monitoring point and the second monitoring point are any two monitoring points with a connection relationship in the acquired initial graph structure. The interval distance between the two monitoring points is calculated based on the position information of the monitoring points with time-invariant characteristics, and then the interval distance and are compared to determine whether to retain the connection relationship between the monitoring points. The distance threshold can be a few meters or tens of meters, and technical personnel in this field can set it according to their own needs.

[0084] Based on this, any two monitoring points with a connection relationship in the initial graph structure are determined, the location information of the two monitoring points is obtained, and the interval distance between the two monitoring points is calculated according to the location information. If the interval distance is less than the preset distance threshold, the edge between the two monitoring points is retained, that is, the connection relationship between the two monitoring points is retained. If the interval distance is not less than (greater than or equal to) the distance threshold, the edge between the two monitoring points is removed, that is, the connection relationship between the two monitoring points is discarded.

[0085] Furthermore, the position representation of the monitoring point can be three-dimensional coordinates, two-dimensional coordinates, or any position representation method. Technical personnel in this field can use different calculation methods to determine the distance between two monitoring points based on different position representation methods.

[0086] Taking the monitoring point coordinates as three-dimensional coordinates as an example, assuming that the first monitoring point and the second monitoring point are monitoring point i and monitoring point j respectively, then the monitoring point coordinates are (xi ,y i ,z i ) and (x j ,y j ,z j ), the distance between monitoring points is calculated as follows:

[0087]

[0088] Set the monitoring point distance relationship threshold to , and retain the connection relationship between monitoring points whose distance is lower than the threshold. That is, if , then retain the connection relationship between the monitoring point and , otherwise, discard it.

[0089] In summary, determining whether to simplify the connection relationship of the monitoring points by the interval distance of the monitoring points and removing redundant connections formed due to being too far away in space can make the key geographical relationships and interaction patterns more prominent, making it easier for the subsequent neural network to accurately analyze the actual impact and dependency relationship between the monitoring points on the slope.

[0090] It should be noted that in the process of simplifying the connection relationship, if the monitoring points are arranged reasonably and densely, there will be no problem of scattered points. If scattered points appear, the connection relationship between the scattered points and the nearest monitoring point can be retained, and the distance threshold can be used as the weight between the two monitoring points.

[0091] In addition to the above method of simplifying the connection relationship, the connection relationship of the initial graph structure can also be simplified based on the displacement of the monitoring point.

[0092] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the method of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure further includes:

[0093] Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0094] Obtaining the cumulative displacement of the first monitoring point and the second monitoring point within a preset time and the displacement at the current moment;

[0095] Determine the displacement connection strength between the first monitoring point and the second monitoring point according to the displacement accumulation amount and the displacement amount at the current moment;

[0096] When the displacement connection strength is less than a connection strength threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0097] When the displacement connection strength is not less than a connection strength threshold, the edge between the first monitoring point and the second monitoring point is removed.

[0098] Specifically, there may be multiple initial graph structures, which are a sequence, wherein each initial graph structure includes each monitoring point in the monitored area at the current moment and the connection relationship between each monitoring point, and the initial graph structure sequence includes each monitoring point at multiple moments and their connection relationship.

[0099] In specific implementation, the time-invariant feature includes the displacement of the monitoring point at the current moment relative to the previous moment, which can be directly obtained. Of course, the time-invariant feature can only include the position information of the monitoring point, and the displacement can be obtained based on the position information of the monitoring point at the current moment and the position information of the monitoring point at the previous moment. The displacement of the monitoring point in the time-invariant feature within the preset time is obtained, and the cumulative displacement of the monitoring point within the preset time is obtained according to the displacement within the preset time. The displacement connection strength of the two monitoring points is calculated according to the obtained cumulative displacement of the two monitoring points and the displacement at the current moment. According to a pre-set connection strength threshold, the connection relationship of the monitoring points is retained or truncated, and the connection relationship of the monitoring points whose distance is lower than the threshold is retained, that is, if the displacement connection strength is less than the connection strength threshold, the connection relationship between the two monitoring points is retained, otherwise, it is discarded.

[0100] Specifically, the preset time is T, and the cumulative displacement of the monitoring point within a period of time T is M T :

[0101]

[0102] Among them, m t is the displacement of each monitoring point in the time period from t-1 to t.

[0103] The displacement connection strength of the two monitoring points is calculated by the following formula:

[0104]

[0105] Among them, x ik 、x jk are the displacements of monitoring point i and monitoring point j at the current moment, are the cumulative displacements of monitoring point i and monitoring point j respectively.

[0106] Set the connection strength threshold eps con , the connection relationship of the monitoring points is retained or truncated, and the connection relationship of the monitoring points whose distance is lower than the threshold is retained. That is, if s ij <eps con , then the connection relationship between monitoring points i and j is retained, otherwise it is discarded.

[0107] In addition to the above two methods of simplifying connection relationships, the connection relationships of the initial graph structure can also be simplified based on the adjacency matrix of the monitoring points and the spatial diffusion intensity.

[0108] In one or more embodiments of the present specification, the connection relationship of the monitoring points in the initial graph structure is represented by edges; and the method of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure further includes:

[0109] Determine the adjacency matrix between each monitoring point;

[0110] Obtaining a transfer matrix between each monitoring point based on the adjacency matrix;

[0111] Determine the displacement characteristic diffusion matrix between the monitoring points according to the transfer matrix and the spatial diffusion intensity;

[0112] Determine a first monitoring point and a second monitoring point, and determine target displacement feature diffusion values ​​corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure;

[0113] When the target displacement feature diffusion value is less than a preset spatial diffusion threshold, retaining the edge between the first monitoring point and the second monitoring point;

[0114] When the target displacement feature diffusion value is not less than a preset spatial diffusion threshold, an edge between the first monitoring point and the second monitoring point is removed.

[0115] Specifically, the adjacency matrix of the monitoring point is obtained according to the matrix of the monitoring point and the unit matrix, the degree matrix of the monitoring point is determined according to the adjacency matrix of the monitoring point, and then the diagonal matrix corresponding to the degree matrix is ​​obtained according to the degree matrix of the monitoring point, the displacement transfer matrix of the monitoring point is determined based on the adjacency matrix, the degree matrix of the monitoring point and the diagonal matrix corresponding to the degree matrix, the displacement feature diffusion matrix is ​​calculated according to the displacement transfer matrix and the spatial diffusion intensity, and the displacement feature diffusion matrix is ​​compared with a preset spatial diffusion threshold. When the displacement feature diffusion matrix is ​​less than the preset spatial diffusion threshold, the edge between the first monitoring point and the second monitoring point is retained, that is, the connection relationship between the two monitoring points is retained. When the displacement feature diffusion matrix is ​​not less than the preset spatial diffusion threshold, the edge between the first monitoring point and the second monitoring point is removed, that is, the connection relationship between the two monitoring points is discarded.

[0116] In specific implementation, the following formula can be used for calculation:

[0117] A loop =I N +A

[0118] D loop =diag(A loop 1)

[0119]

[0120] Among them, A loop is an adjacency matrix with self-loops, D loop represents the degree matrix of the node, is a diagonal matrix, Each element in D loop The inverse square root of the corresponding element.

[0121] Then the displacement characteristic diffusion matrix is ​​calculated as follows:

[0122] D S =∝ S ×(I N -(1-∝ S )×T sym ) -1

[0123] Among them, ∝ S is the spatial diffusion intensity, I N is the identity matrix of dimension N×N.

[0124] According to the set spatial diffusion threshold eps of the time-invariant characteristics between monitoring points sdis , that is, the maximum value of the spatial relationship between monitoring points, retaining the connection relationship of monitoring points whose distance is lower than this threshold. ij <eps sdis , then the connection relationship between monitoring points i and j is retained, where D ij is the calculated displacement characteristic diffusion matrix D S The element value in the i-th row and j-th column is replaced by the value in the j-th column. Otherwise, it is discarded.

[0125] In summary, through the above steps, we can not only quantify the propagation effect of the displacement of the monitoring points in space, but also effectively filter out insignificant connections by setting thresholds, simplify the graph structure, and focus more on the key connections that have a significant impact on slope stability.

[0126] Considering that the time-varying characteristic data such as slope internal deformation, stress, groundwater level, blasting vibration, and rainfall data have an impact on the future displacement of the slope, an attenuation function can be established to determine the impact of the time-varying characteristics on the future displacement of the slope.

[0127] In one or more embodiments of the present specification, after simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant features of the monitoring points to obtain the intermediate graph structure, the method further includes:

[0128] Obtaining a preset time lag range and a preset time lag intensity;

[0129] determining a diffusion intensity variation function according to the time lag range and the time lag intensity;

[0130] The intermediate graph structure data is subjected to time diffusion convolution according to the diffusion intensity variation function.

[0131] Specifically, the time lag range of the time variation characteristics at the same monitoring point is set to T lag , let the time diffusion intensity, that is, the time characteristic lag intensity at the same monitoring point, be ∝ T ;

[0132] Time lag range T lag The physical meaning of is the time range of the past time variation characteristics that affect the future displacement, and the time characteristic lag intensity ∝ T The physical meaning of is the degree of influence of past time variation characteristics on future displacement. This influence usually weakens with the increase of time interval, that is, the earlier the time, the smaller the influence on the future. This attenuation effect can be modeled by various functions.

[0133] Preferably, we use an exponential decay function to represent the change in diffusion intensity:

[0134]

[0135] Among them, λ is the decay exponent and t is the current time.

[0136] Perform time-diffused convolution on the time-varying features of the monitoring points;

[0137] Specifically, the time-varying characteristic sequence within the time lag range is assumed to be Then the calculation method for performing time-diffused convolution is as follows:

[0138]

[0139] By adjusting the decay coefficient λ, the decay rate of the influence of past features on current features can be controlled.

[0140] like Figure 3 As shown, the displacement (or other characteristic value) at each time point in the past is calculated and multiplied by the influence weight (given by an exponential decay function) under the corresponding time lag, and then all these products are added together to obtain the "diffusion influence" value of the current position.

[0141] Step 106: Analyze the spatial features of the intermediate graph structure using a graph neural network to obtain spatial information, and analyze the spatial information and the time information of the previous moment using a recurrent neural network to obtain the time information of the current moment.

[0142] Specifically, using a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information specifically refers to updating the attribute information of the monitoring point according to a set of monitoring points that have a connection relationship with the monitoring point, capturing the influence of other monitoring points on the monitoring point, and thus analyzing to obtain the spatial information of the intermediate graph structure, that is, using a graph neural network to capture the spatial dependency between monitoring points; using a recurrent neural network (RNNs) to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment, specifically refers to capturing the time dependency between different time slices according to the influence of the monitoring points within the time lag range on the monitoring points at the current moment.

[0143] In specific implementation, the graph neural network processes spatial relationships in the following way:

[0144]

[0145] Among them, N(i) is the neighbor set of node i, and GNN is the aggregation function of graph neural network.

[0146] The output of RNN at each time slice It is expressed as:

[0147]

[0148] in, is the hidden state of the previous moment, and RNN is the update function of the recurrent neural network.

[0149] Step 108: Input the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment.

[0150] In specific implementation, the regressor is used to obtain the final displacement prediction:

[0151]

[0152] Wherein, Regressor is a regressor, and preferably, a neural network or other regression models are used.

[0153] In summary, the slope displacement prediction method of this specification reduces the connection density of the monitoring point graph by simplifying the connection relationship of the monitoring points, and uses graph diffusion and time diffusion methods to perform diffusion modeling on the time-invariant characteristics and time-varying characteristics of the monitoring points, captures the time lag effect of the time-varying characteristics on the displacement, and captures its spatiotemporal relationship through a spatiotemporal model to obtain the displacement prediction result of the model. It can not only efficiently process large-scale spatial data, but also accurately capture and predict the complex dynamic relationships in time series data, providing strong technical support for the safety monitoring of open-pit coal mine slopes.

[0154] The following combination Figure 4 , taking the application of the slope displacement prediction method provided in this specification in slope displacement prediction as an example, the slope displacement prediction method is further described. Figure 4 A processing flow chart of a slope displacement prediction method provided in an embodiment of the present specification is shown, which specifically includes the following steps.

[0155] Step 402: Acquire an initial graph structure of the area to be detected at the current moment, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points.

[0156] Step 404: Simplify the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure.

[0157] The connection relationship of the monitoring points in the initial graph structure can be simplified by the following method:

[0158] (1) determining a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in an initial graph structure; obtaining position information of the first monitoring point and the second monitoring point, and determining a distance between the first monitoring point and the second monitoring point according to the position information; retaining an edge between the first monitoring point and the second monitoring point when the distance between the first monitoring point and the second monitoring point is less than a distance threshold; and removing an edge between the first monitoring point and the second monitoring point when the distance between the first monitoring point and the second monitoring point is not less than the distance threshold.

[0159] (2) Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure. Obtain the displacement accumulation of the first monitoring point and the second monitoring point within a preset time and the displacement at a target time. Determine the displacement connection strength of the first monitoring point and the second monitoring point according to the displacement accumulation and the displacement at the target time. If the displacement connection strength is less than a preset connection threshold, retain the edge between the first monitoring point and the second monitoring point. If the displacement connection strength is not less than a preset connection threshold, remove the edge between the first monitoring point and the second monitoring point.

[0160] (3) determining an adjacency matrix between each monitoring point; obtaining a transfer matrix between each monitoring point based on the adjacency matrix; determining a displacement feature diffusion matrix between each monitoring point based on the transfer matrix and the spatial diffusion intensity; determining a first monitoring point and a second monitoring point, and determining a target displacement feature diffusion value corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; retaining an edge between the first monitoring point and the second monitoring point when the target displacement feature diffusion value is less than a preset spatial diffusion threshold; and removing an edge between the first monitoring point and the second monitoring point when the target displacement feature diffusion value is not less than a preset spatial diffusion threshold.

[0161] Step 406: Obtain a preset time lag range and a preset time lag strength.

[0162] Step 408: Determine a diffusion intensity variation function according to the time lag range and the time lag intensity.

[0163] Step 410: Perform time diffusion convolution on the intermediate graph structure data according to the diffusion intensity variation function.

[0164] Step 412: Analyze the spatial features of the intermediate graph structure using a graph neural network to obtain spatial information, and analyze the spatial information and the time information of the previous moment using a recurrent neural network to obtain the time information of the current moment.

[0165] Specifically, the spatial information is obtained by analyzing the spatial features of the intermediate graph structure using a graph neural network, including: obtaining the target monitoring point and the neighbor set of the target monitoring point; aggregating the spatial features of the target monitoring point and the neighbor set of the target monitoring point using the graph neural network to obtain the spatial information of the target monitoring point at the current moment. The spatial information and the time information of the previous moment are analyzed using a recurrent neural network to obtain the time information of the current moment, including: obtaining the time information of the monitoring point at the previous moment; updating the spatial information of the target monitoring point at the current moment and the time information of the monitoring point at the previous moment using the update function of the recurrent neural network to obtain the time information of the current moment.

[0166] Step 414: input the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment.

[0167] In summary, the slope displacement prediction method of this specification reduces the connection density of the monitoring point graph by simplifying the connection relationship of the monitoring points, and uses graph diffusion and time diffusion methods to perform diffusion modeling on the time-invariant characteristics and time-varying characteristics of the monitoring points, captures the time lag effect of the time-varying characteristics on the displacement, and captures its spatiotemporal relationship through a spatiotemporal model to obtain the displacement prediction result of the model. It can not only efficiently process large-scale spatial data, but also accurately capture and predict the complex dynamic relationships in time series data, providing strong technical support for the safety monitoring of open-pit coal mine slopes.

[0168] Corresponding to the above method embodiment, this specification also provides a slope displacement prediction device embodiment, Figure 5 FIG. 2 shows a schematic diagram of a slope displacement prediction device provided by an embodiment of the present specification. Figure 5 As shown, the device comprises:

[0169] An acquisition module 502 is configured to acquire an initial graph structure of the area to be detected at a current moment, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points;

[0170] An intermediate graph structure obtaining module 504 is configured to simplify the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure;

[0171] The time information acquisition module 506 is configured to use a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information, and use a recurrent neural network to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment;

[0172] The displacement prediction result obtaining module 508 is configured to input the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment.

[0173] In one or more embodiments of the present specification, the intermediate graph structure acquisition module is further configured to determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; obtain location information of the first monitoring point and the second monitoring point, and determine an interval distance between the first monitoring point and the second monitoring point based on the location information; retain an edge between the first monitoring point and the second monitoring point when the interval distance is less than a distance threshold; and remove an edge between the first monitoring point and the second monitoring point when the interval distance is not less than a distance threshold.

[0174] In one or more embodiments of the present specification, the intermediate graph structure acquisition module is further configured to determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure. Obtain the cumulative displacement of the first monitoring point and the second monitoring point within a preset time and the displacement at the target moment. Determine the displacement connection strength of the first monitoring point and the second monitoring point based on the cumulative displacement and the displacement at the target moment. When the displacement connection strength is less than a preset connection threshold, retain the edge between the first monitoring point and the second monitoring point. When the displacement connection strength is not less than a preset connection threshold, remove the edge between the first monitoring point and the second monitoring point.

[0175] In one or more embodiments of the present specification, the intermediate graph structure acquisition module is further configured to determine an adjacency matrix between each monitoring point; obtain a transfer matrix between each monitoring point based on the adjacency matrix; determine a displacement feature diffusion matrix between each monitoring point based on the transfer matrix and the spatial diffusion intensity; determine a first monitoring point and a second monitoring point, and determine a target displacement feature diffusion value corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; when the target displacement feature diffusion value is less than a preset spatial diffusion threshold, retain the edge between the first monitoring point and the second monitoring point; when the target displacement feature diffusion value is not less than the preset spatial diffusion threshold, remove the edge between the first monitoring point and the second monitoring point.

[0176] In one or more embodiments of the present specification, the device also includes a time diffusion convolution module, and is also configured to obtain a preset time lag range and a preset time lag strength; determine a diffusion intensity change function based on the time lag range and the time lag strength; and perform time diffusion convolution on the intermediate graph structure data based on the diffusion intensity change function.

[0177] In one or more embodiments of the present specification, the time information acquisition module is also configured to obtain the target monitoring point and the neighbor set of the target monitoring point; use the graph neural network to aggregate the spatial features of the target monitoring point and the neighbor set of the target monitoring point to obtain the spatial information of the target monitoring point at the current moment.

[0178] In one or more embodiments of the present specification, the time information acquisition module is also configured to obtain the time information of the monitoring point at the previous moment; use the update function of the recurrent neural network to update the spatial information of the target monitoring point at the current moment and the time information of the monitoring point at the previous moment to obtain the time information at the current moment.

[0179] In summary, the slope displacement prediction device of the present specification reduces the connection density of the monitoring point graph by simplifying the connection relationship of the monitoring points, and uses graph diffusion and time diffusion methods to perform diffusion modeling on the time-invariant characteristics and time-varying characteristics of the monitoring points, captures the time lag effect of the time-varying characteristics on the displacement, and captures its spatiotemporal relationship through a spatiotemporal model to obtain the displacement prediction result of the model. It can not only efficiently process large-scale spatial data, but also accurately capture and predict the complex dynamic relationships in time series data, providing strong technical support for the safety monitoring of open-pit coal mine slopes.

[0180] The above is a schematic scheme of a slope displacement prediction device of this embodiment. It should be noted that the technical scheme of the slope displacement prediction device and the technical scheme of the above-mentioned slope displacement prediction method belong to the same concept, and the details not described in detail in the technical scheme of the slope displacement prediction device can be referred to the description of the technical scheme of the above-mentioned slope displacement prediction method.

[0181] Figure 6 The block diagram of a computing device 600 according to an embodiment of the present specification is shown. The components of the computing device 600 include but are not limited to a memory 610 and a processor 620. The processor 620 is connected to the memory 610 via a bus 630, and the database 650 is used to store data.

[0182] The computing device 600 also includes an access device 640 that enables the computing device 600 to communicate via one or more networks 660. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 640 may include one or more of any type of network interface (e.g., a network interface card (NIC)) of wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a world-wide interoperability for microwave access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, and a near field communication (NFC).

[0183] In one embodiment of the present specification, the above components of the computing device 600 and Figure 6 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 6 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0184] The computing device 600 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or a personal computer (PC). The computing device 600 may also be a mobile or stationary server.

[0185] The processor 620 is used to execute the following computer executable instructions, which, when executed by the processor, implement the steps of the above slope displacement prediction method.

[0186] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above slope displacement prediction method belong to the same concept, and the details not described in detail in the technical scheme of the computing device can be referred to the description of the technical scheme of the above slope displacement prediction method.

[0187] An embodiment of the present specification further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the above-mentioned slope displacement prediction method.

[0188] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned slope displacement prediction method belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the above-mentioned slope displacement prediction method.

[0189] An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned slope displacement prediction method.

[0190] The above is a schematic scheme of a computer program of this embodiment. It should be noted that the technical scheme of the computer program and the technical scheme of the above slope displacement prediction method belong to the same concept, and the details not described in detail in the technical scheme of the computer program can be found in the description of the technical scheme of the above slope displacement prediction method.

[0191] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0192] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0193] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.

[0194] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0195] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to only the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can well understand and use this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. A slope displacement prediction method, characterized in that: include: Acquire an initial graph structure of the area to be monitored before the current time, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points; Simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure; Analyzing the spatial features of the intermediate graph structure using a graph neural network to obtain spatial information, and analyzing the spatial information and time information at a previous moment using a recurrent neural network to obtain time information at a current moment; Inputting the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment; Wherein, the connection relationship of the monitoring points in the initial graph structure is represented by an edge; the connection relationship of the monitoring points in the initial graph structure is simplified based on the time-invariant characteristics of the monitoring points to obtain the intermediate graph structure, including: determining a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; obtaining the cumulative displacement of the first monitoring point and the second monitoring point within a preset time and the displacement at the current moment; determining the displacement connection strength of the first monitoring point and the second monitoring point according to the cumulative displacement and the displacement at the current moment; when the displacement connection strength is less than a connection strength threshold, retaining the edge between the first monitoring point and the second monitoring point; when the displacement connection strength is not less than a connection strength threshold, removing the edge between the first monitoring point and the second monitoring point; and / or, Determine an adjacency matrix between each monitoring point; obtain a transfer matrix between each monitoring point based on the adjacency matrix; determine a displacement feature diffusion matrix between each monitoring point according to the transfer matrix and the spatial diffusion intensity; determine a first monitoring point and a second monitoring point, and determine a target displacement feature diffusion value corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; when the target displacement feature diffusion value is less than a preset spatial diffusion threshold, retain the edge between the first monitoring point and the second monitoring point; when the target displacement feature diffusion value is not less than the preset spatial diffusion threshold, remove the edge between the first monitoring point and the second monitoring point; After simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain the intermediate graph structure, the method also includes: obtaining a preset time lag range and a preset time lag strength; determining a diffusion intensity change function according to the time lag range and the time lag strength; and performing time diffusion convolution on the intermediate graph structure data according to the diffusion intensity change function.

2. The slope displacement prediction method according to claim 1, characterized in that: The step of simplifying the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure further includes: Determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; Acquire location information of the first monitoring point and the second monitoring point, and determine the interval distance between the first monitoring point and the second monitoring point according to the location information; When the interval distance is less than a distance threshold, retaining the edge between the first monitoring point and the second monitoring point; When the interval distance is not less than a distance threshold, the edge between the first monitoring point and the second monitoring point is removed.

3. The slope displacement prediction method according to claim 1, characterized in that: The using of a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information includes: Obtaining a target monitoring point and a neighbor set of the target monitoring point; The graph neural network is used to aggregate the spatial features of the target monitoring point and the neighbor set of the target monitoring point to obtain the spatial information of the target monitoring point at the current moment.

4. The slope displacement prediction method according to claim 3, characterized in that: The using of a recurrent neural network to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment includes: Obtaining spatial information of the monitoring point at the current moment; The updating function of the recurrent neural network is used to update the spatial information of the target monitoring point at the current moment and the time information of the monitoring point at the previous moment to obtain the time information at the current moment.

5. A slope displacement prediction device, characterized in that: include: An acquisition module is configured to acquire an initial graph structure of the area to be monitored at a current moment, wherein the initial graph structure includes at least two monitoring points and time-invariant features of the monitoring points; an intermediate graph structure obtaining module, configured to simplify the connection relationship of the monitoring points in the initial graph structure based on the time-invariant characteristics of the monitoring points to obtain an intermediate graph structure; The intermediate graph structure acquisition module is further configured to determine a first monitoring point and a second monitoring point, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; obtain the displacement accumulation amount of the first monitoring point and the second monitoring point within a preset time and the displacement amount at the current moment; determine the displacement connection strength of the first monitoring point and the second monitoring point according to the displacement accumulation amount and the displacement amount at the current moment; if the displacement connection strength is less than a connection strength threshold, retain the edge between the first monitoring point and the second monitoring point; if the displacement connection strength is not less than the connection strength threshold, remove the edge between the first monitoring point and the second monitoring point; And / or, the intermediate graph structure acquisition module is further configured to determine an adjacency matrix between each monitoring point; obtain a transfer matrix between each monitoring point based on the adjacency matrix; determine a displacement feature diffusion matrix between each monitoring point according to the transfer matrix and the spatial diffusion intensity; determine a first monitoring point and a second monitoring point, and determine a target displacement feature diffusion value corresponding to the first monitoring point and the second monitoring point based on the displacement feature diffusion matrix, wherein the first monitoring point and the second monitoring point are any two monitoring points in the initial graph structure; when the target displacement feature diffusion value is less than a preset spatial diffusion threshold, retain the edge between the first monitoring point and the second monitoring point; when the target displacement feature diffusion value is not less than the preset spatial diffusion threshold, remove the edge between the first monitoring point and the second monitoring point; A time diffusion convolution module is configured to obtain a preset time lag range and a preset time lag strength; determine a diffusion strength variation function according to the time lag range and the time lag strength; and perform time diffusion convolution on the intermediate graph structure data according to the diffusion strength variation function; A time information acquisition module is configured to use a graph neural network to analyze the spatial features of the intermediate graph structure to obtain spatial information, and use a recurrent neural network to analyze the spatial information and the time information of the previous moment to obtain the time information of the current moment; The displacement prediction result obtaining module is configured to input the time information of the current moment into the regressor to obtain the displacement prediction result of the current moment.

6. A computing device, characterized in that include: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the slope displacement prediction method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium, characterized in that: It stores computer executable instructions, which, when executed by a processor, can implement the steps of the slope displacement prediction method described in any one of claims 1 to 4.

8. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the slope displacement prediction method according to any one of claims 1 to 4.

Images